——不要太惊讶，是让9子的前提下。。。。。。。

800个CPU的超级电脑首次战胜职业围棋棋手

ugmbbc发布于 2008-08-09 09:06:07|4386 次阅读 字体：大 小 打印预览


感谢everypot.com的投递
新闻来源:slashdot.org
昨天，在美国围棋大会上，名为MoGo的超级计算机在被让9子的对局中以微弱优战胜了一位职业8段棋手。MoGo有800 个运行在4.7GHz 的CPU，拥有15 Teraflops的浮点运算能力，运行的软件基于Monte Carlo 算法。



访问:Computer Beats Pro At US Go Congress
http://games.slashdot.org/games/08/08/08/1243244.shtml

棋谱：
http://www.usgo.org/congress/2008/myungwan-mogo.sgf
(;GM[1]FF[4]CA[UTF-8]AP[CGoban:3]ST[2]
RU[Chinese]SZ[19]HA[9]KM[0.50]TM[3600]
PW[MyungWan]PB[MoGoTiTan]DT[2008-08-07]PC[The KGS Go Server at http://www.gokgs.com/]C[MoGoTiTan [-\]: GTP Engine for MoGoTiTan (black): MoGo version 3.0.0: I'll resign if you win and playing until you pass if I win. If you passed and I did not, it often means that the status of some stones are not clear enough for me (I want to be sure we will agree on dead stones). Clarify them. Maybe I am losing, but have not enough time to be sure I should resign. In this case play until the end :'( (very rare cases)
]RE[B+1.50]
;B[dd]BL[3599.305]
;B[dj]BL[3599.013]
;B[dp]BL[3598.82]
;B[pd]BL[3598.599]
;B[pj]BL[3598.385]
;B[pp]BL[3598.168]
;B[jd]BL[3597.954]
;B[jp]BL[3597.739]
;B[jj]BL[3597.528]C[MyungWan [-\]: hi
Fatality [7k\]: hi lol
]
;W[pg]WL[3588.492]C[Francisco [3d\]: lol
ahmadr [10k\]: Gl
ibd [16k\]: dun dun dunnnnn
teytaud [-\]: Hi
NPcomplete [5k\]: hi master :)
certo [1k\]: no handicap!?
Francisco [3d\]: he is very nervous
certo [1k\]: well besides 9
trumpet [3k\]: no handi hehe
sixsix [-\]: lol
]
;B[md]BL[3581.62]
;W[ql]WL[3583.335]C[pasky [1k\]: I will be tracking this game with CzechBot, running on dual Intel Xeon on 2.4GHz, and give some details on what my mogo is thinking about the moves.
Francisco [3d\]: I am strongly betting on MogoTitan
NPcomplete [5k\]: me too
trumpet [3k\]: great pasky
]
;B[mj]BL[3564.556]C[PowerP [3d\]: honestly i would have a hard time taking this seriously if i was myung
Yuura [8k\]: et c'est reparti pour une bastonade XD
]
;W[nj]WL[3575.703]C[Bovido [8k\]: Go MoGo!
certo [1k\]: mogo dead, 9h is not enough
danteVAM [5k\]: have a good game again =)
ekku [-\]: what is b running on?
CrazyApple [2k\]: 1h
kuikentje [5k\]: I'm strongly betting on MyungWan
joelw [12k\]: titan is computer?
ketchup [6k\]: certo, we'll see, if the pro wanted to play him on 9h
hikaru78 [15k\]: go both!
]
;B[ni]BL[3546.864]C[ahmadr [10k\]: 9h + 4x time
ketchup [6k\]: he had good reason to
Hu [3k\]: This is a much longer time setting.
trumpet [3k\]: being the pro you can only loose here
hikaru78 [15k\]: o.o
]
;W[oi]WL[3572.78]C[sixsix [-\]: i have it cloned in case anything worth discussing comes up
zeb [8k\]: so exciting
jhouse [3k\]: It'd be interesting to see mogo have 12 stones for this game :)
lukaszchol [13k?\]: I bet on MOGop this time
McCrusty [1k\]: what is w running on? green tea?
Ho [5k\]: did they finnish this last USGO game after lunch?
]
;B[nk]BL[3528.301]
;W[oj]WL[3570.841]C[Almdudler [3d\]: nyu
Tsuijin [1d\]: white wants to finish off mogo fast and get back to other game
ketchup [6k\]: since myungwan is playing in it
Sakae [-\]: maybe hte pro let the bot win on purpose?
ketchup [6k\]: i highly doubt it
Sakae [-\]: to cheer us up :o
IanO: what is the playout rate at this time control?
]
;B[ok]BL[3508.822]C[ketchup [6k\]: it's supposed to take place after this game, Ho
]
;W[mi]WL[3561.525]C[Batavia [5k\]: so what kind of spec does mogo run on?
Ho [5k\]: thx
]
;B[pi]BL[3486.931]
;W[oh]WL[3558.789]C[Kashuu [1k\]: yes im also curious
ketchup [6k\]: r8 alive
Kashuu [1k\]: is mogo playing at a super computer now?
ekku [-\]: what kind of hardware?
toxygen [?\]: 800 cores
]
;B[pk]BL[3465.462]C[toxygen [?\]: xeons i guess, but i'm not sure
]
;W[lj]WL[3555.843]C[Sebass [7k\]: XD
TanGeng [2d?\]: 800 cores?
Kashuu [1k\]: only 800 cores?
toxygen [?\]: yep
chrono3450 [1d\]: he's playing better, that's for sure
chrono3450 [1d\]: thank god
ekku [-\]: that's quite a heap of flops :D
ketchup [6k\]: myungwan still playing so fast..
Tsuijin [1d\]: its cause he knows it is over
]
;B[mk]BL[3443.222]C[Sebass [7k\]: Wait, is MoGo a computer player?
]
;W[lh]WL[3553.134]C[toxygen [?\]: yes
sixsix [-\]: lol
TanGeng [2d?\]: they don't use flops
Hu [3k\]: Someone said 800 processor x quadcores = about 3000.
AdrianG430 [12k\]: -_-
Sebass [7k\]: LOL
Amano [12k\]: yes
Sebass [7k\]: Nice.
hikaru78 [15k\]: lol
iphicles [1k?\]: Playing fast gives computer less time to think
toxygen [?\]: Hu: not true apparently
Unwichtig [18k\]: short
TanGeng [2d?\]: only fixed point math
]
;B[lk]BL[3419.983]C[Singer [10k\]: Where did he get 800 processors from?
]
;W[qp]WL[3548.748]C[kuikentje [5k\]: I saw the same thing Hu
navilus [1k\]: lol
Celila [-\]: hu apparently it's much less .. it was meant to be like that, but it's not
Elandal [2k\]: nodes, cores, whatever. It's better than my computer.
tsiki [1d\]: only 9 handi?
ibd [16k\]: what's the benefit of only fixed point on today's hardware?
Sakae [-\]: eheh I saw a Hu Bin on EGC
Candidate [2k?\]: i could have provided 100 cores is i had known this was coming up... ^^
toxygen [?\]: singer: i was told that these information will be disclosed after match
]
;B[qo]BL[3395.968]C[royalty [3d\]: well, it's not doin bad yet
Celila [-\]: it's 1 h thinking time now, tsiki
]
;W[po]WL[3546.312]C[TanGeng [2d?\]: they key si the algorithm
ketchup [6k\]: celila what is it now then?
tsiki [1d\]: yes but i thought the previous games would be used to determine the handi
Sakae [-\]: this one doesn't have an algorithim..
pasky [1k\]: so far they play too fast to really give me a chance to follow them :)
KintaroOe [5k\]: wait so mogo is a comp.?
gentil [6d\]: how strong is mogotitan ?
]
;B[qq]BL[3370.932]C[Sakae [-\]: it thinkgs by reading random moves
]
;W[rp]WL[3543.568]C[tsiki [1d\]: i guess they think MoGo gains a lot more from the extra time
Sebass [7k\]: yup.
iphicles [1k?\]: We are finding out
Celila [-\]: 800 apparently
ekku [-\]: sakae, surely it does
Sakae [-\]: and taking the percentage..
TanGeng [2d?\]: that's an algorithm
Kashuu [1k\]: the key is.. a computer should be able to learn from its mistakes and improve
Tsuijin [1d\]: if pro loses will the play another?
trumpet [3k\]: mogo is a software for being exact
progers [12k\]: Kim should gain a lot from the time too
ketchup [6k\]: kashuu.. that doesnt work..
yoyoma [1d\]: oh please mogo don't hose this corner up
]
;B[pn]BL[3344.997]
;W[oo]WL[3541.21]C[telchar [-\]: a 9*9 game against MyungWan would have been interesting too :)
TwoStoned: mogo has a good chance, if he stays away from the sake
Amano [12k\]: ^^
PowerP [3d\]: these hanes are not good
Hagios [4k\]: Well, it played this corner differently than the last two times...
jhouse [3k\]: strength for mogotitan is probably very time dependent. It has less memorized knowledge, so gets weak with short time intervals
SadShape [1k\]: we will see if mogo make the same joseki here...
]
;B[rq]BL[3318.382]C[NemoEimi [?\]: a 9x9 game would have been VERY interesting in my book :)
jsc [14k\]: proof that score estimate only works in the endgame...........
Batavia [5k\]: anywhere/anyone who could give some information about what the usage is those 800 processors?
alexanro [7k\]: score estimate works fine after the first move as well
Zyphe [9k\]: hmmm wouldnt you say that 9 or even 12 stone handi isnt a true estimate of mogo's strength due to the fact that every player is different in their proficiency with handicap go?
pebarefoot [12k\]: sake no good for electronics ! the potched watt never toils
joelw [12k\]: it's hard for a 9x9 game to be very interesting if there's handi involved
]
;W[ro]WL[3506.423]C[Kashuu [1k\]: at work i sometimes play gnugo and give him 9 stones, ive played the first 89 moves in the exact same order.. after he lost to me with the same moves last round, what im trying to say it.. computers dont seem to learn
kuikentje [5k\]: if I were to estimate who were ahead - without knowledge of the players - I'd say black too, jsc
Sebass [7k\]: Batavia: Wikipedia.
ekku [-\]: i think you could tweak the monte carlo with heuristics from analyzer pro games
Celila [-\]: mogo is simulating possible scenarios, it's not using pro games
ekku [-\]: so that it seeds more likely paths to the elimination
gentil [6d\]: next is q3
]
;B[qn]BL[3290.811]C[gentil [6d\]: and capture somthing
TanGeng [2d?\]: ahh the monte carlo method
]
;W[pq]WL[3503.865]C[NemoEimi [?\]: 9x9 should of course be played without handicap :)
Amano [12k\]: ^^
Kashuu [1k\]: there are not alot of pro games anyway with 9 stones ;p
Tors [5k\]: does mogo think during opponents turn?
ibd [16k\]: Kashuu: gnugo is a program that responds to same moves the same, always. mogo shouldn't do that
TanGeng [2d?\]: nice random walk algorithms
NemoEimi [?\]: mogo is not yet at the point of being able to give a pro handicap at 9x9 ;)
jtn [6k\]: tors: yes
teytaud [-\]: Tors: yes.
]
;B[op]BL[3262.202]
;W[pr]WL[3502.053]C[gentil [6d\]: o5
McCrusty [1k\]: Myungwan's goal is to win using the least time possible?
Tors [5k\]: thanks
Dragonette [6k\]: MyungWan has made 20 moves in 1 min 38 sec, mogo ain't getting much thinking time
gentil [6d\]: if b is strong, o5
Zyphe [9k\]: computer go will take off when we develope more complex learning ai!
zpt [2d?\]: USGO, is it possible to arrange 9x9 game after this game?
Singer [10k\]: Is it generally recognised that the pro would completely pwn the bot without a handicap?
alexanro [7k\]: now mogo plays r2?
Hagios [4k\]: Far as I'm aware, mogo only beat a pro once, as white, with 7.5 komi on 9x9
]
;B[no]BL[3232.706]C[XY0908 [-\]: Mr. Kim is using another computer.
]
;W[rn]WL[3498.865]C[gentil [6d\]: damm
ketchup [6k\]: zpt, i think myungwan still has a game to play
chrono3450 [1d\]: B got cheated here
ketchup [6k\]: for the usopen
szaloony [3k\]: QUIZ: who invented monte carlo algorithm and is the father of hydrogen bomb? (skynet is coming..)
mlnmiles [1d\]: P:)
zpt [2d?\]: ketchup, are you sure?
ketchup [6k\]: yes
IanO: Teller used MC?
mlnmiles [1d\]: NJeumann
TanGeng [2d?\]: yeah, black is over concentrated
]
;B[on]BL[3202.455]C[mlnmiles [1d\]: and Ulam
ato [6k\]: no
ketchup [6k\]: last time i checked..
mlnmiles [1d\]: Neumann* and Ulam
ato [6k\]: father of hydrogen bomb is teller
]
;W[dn]WL[3485.26]C[jhouse [3k\]: was capturing that ladder necessary? I would think not
mlnmiles [1d\]: Openheimer
XapoH [4k\]: 小邪褏邪褉芯胁 %)
ato [6k\]: and MC is by Ulam
blackrain [6k\]: skyneyt is a satt. deployed by england is already up
Squirrel [-\]: mogo wuz robbed
Rakeliem [1d\]: o2
Sakae [-\]: this bot know no sabaki or nothing... it's really...
PowerP [3d\]: b is just a touch overconcentrated
Sebass [7k\]: It wasn't needed.... I think.
szaloony [3k\]: teller is called father of hydrogen bomb. who is the real father?
ato [6k\]: oppenheimer is father of manhattan projest -> atomic bomb
kuikentje [5k\]: it's a bot sakae.
]
;B[qr]BL[3171.223]C[Kashuu [1k\]: I cant shake the feeling off me that this bot is much stronger when you play it without stones
royalty [3d\]: this again
telchar [-\]: overconcentration is acceptable in 9h game...
alexanro [7k\]: hah, knew mogo would play r2 :)
Kashuu [1k\]: atleast the openings look decent then
tromp [-\]: better p3?
pasky [1k\]: I hate git-gui T.T
TanGeng [2d?\]: it's playing very conservatively despite its it being so early in the game
]
;W[nq]WL[3467.314]C[Rakeliem [1d\]: poor mogo
ekku [-\]: it's not trying to emulate human deduction, it's just crunching propabilities.
certo [1k\]: 800 vs 1 is sort of unfair, they should make it 800 pros vs 800 computers
PowerP [3d\]: hahaha
Amano [12k\]: too hard
Nyanjilla [-\]: mogo did that bottom right cornr last time
IanO: 800 pros would never agree
toxygen [?\]: ha
chrono3450 [1d\]: it's not unfair because the computer is not human
Amano [12k\]: imagine 800 pros wondring about where to play
ibd [16k\]: it playing conservatively means it thinks it is winning.
Sebass [7k\]: So, MoGo is using a brute force algorithm?
]
;B[oq]BL[3139.139]
;W[or]WL[3465.198]C[trumpet [3k\]: then the bot might stand a chance
Candidate [2k?\]: certo, the pros would lose on time :D
jhouse [3k\]: 800 computers vs 800 of me = mogo victory
Sakae [-\]: certo, humans doesn't posses that superior teamwork thinking
kuikentje [5k\]: not entirely Sebass
og [9k\]: sebass, wikipedia computer go, all answers are there
Kashuu [1k\]: 800 pro vs 800 cores = bot wins on time
kukulis4 [2d\]: white already got free move at D6
gentil [6d\]: b alive in corner
kuikentje [5k\]: a summary of the algoritm can be found here:
alexanro [7k\]: Nyanjilla: oh, ok :)
kuikentje [5k\]: http://www.lri.fr/~gelly/MoGo.htm
Yaxamie [14k\]: just take a democratic result from the pros
Sebass [7k\]: Thanks :D
Aliens243 [9k?\]: the pros using a kind of voting system to choose the moves
gentil [6d\]: just t2
]
;B[mr]BL[3106.108]C[Sakae [-\]: here's my summary of the algorithim...: Random
]
;W[nr]WL[3461.027]C[royalty [3d\]: democratic result-> average of the coordinates
Candidate [2k?\]: We know! Now shu...
floss [5k\]: there are 10^9 processors in myungwan's case. 3000 is nothing.
mlnmiles [1d\]: not that random
Rakeliem [1d\]: black might live in corner
Rakeliem [1d\]: with ko threat
iphicles [1k?\]: Show us?
Sebass [7k\]: :D
]
;B[mq]BL[3072.213]C[zpt [2d?\]: human processors are much weaker, floss
]
;W[qs]WL[3455.338]C[Amano [12k\]: not sure
bearclaw: floss: but working at, what, 100Hz top each? :)
floss [5k\]: just pointing out the absurdity
DonkeyKong [11k\]: live..how?
Squirrel [-\]: i dont think 800 pros voting would be better unless they discussed
ibd [16k\]: floss: but myungwan can't calculate millions of multiplications a second, for example
ekku [-\]: humans aren't processors
royalty [3d\]: r1?
CHUCHU [9k\]: how can he live?
ekku [-\]: we're deductors
pasky [1k\]: git-gui is just unusable :(
Rakeliem [1d\]: atleast if white tried n4
CHUCHU [9k\]: t2
Sebass [7k\]: Black S1 might allow ko for life?
Aliens243 [9k?\]: a neuron can't be compared to a processor.
Sebass [7k\]: Err
Sebass [7k\]: Nevermind.
kuikentje [5k\]: he also can't read quite as many moves as Mogo - but he knows which moves will end up nowhere in the blink of an eye
chrono3450 [1d\]: millions? someone is dillusional
]
;B[rs]BL[3037.364]C[Aliens243 [9k?\]: a transistor at most
DeCard88 [13k\]: what means 8p ?
]
;W[sq]WL[3452.705]C[kuikentje [5k\]: intuition is the key here
Tsuijin [1d\]: hands up!
Sebass [7k\]: 8-professional.
mrkibbles [1d?\]: interesting
Amano [12k\]: ^p means like :p
Sebass [7k\]: Ahaha
Longshi [-\]: the kgs kibitz is the random factor here
Amano [12k\]: 8 for the eyes
chrono3450 [1d\]: 8p = 8 dan pro
tromp [-\]: w mistake?
Rakeliem [1d\]: yea looks like w dead
pasky [1k\]: er, I mean gogui :)
kilroy [2d\]: isn't the standard tesuji s1 instead of r1?
]
;B[ps]BL[3001.691]
;W[sr]WL[3450.174]C[masek [1d\]: white want ko
kilroy [2d\]: maybe white wants sente?
Zyphe [9k\]: i would disagree with the humans ability to crunch numbers compared to that of a processor
gentil [6d\]: o4
SadShape [1k\]: looks like myungwan will have a very hard job in this game
chrono3450 [1d\]: wow
chrono3450 [1d\]: B kills W
chrono3450 [1d\]: this early
ketchup [6k\]: p1
royalty [3d\]: t4
kukulis4 [2d\]: black O1 now ?
seku [1k\]: zyphe: professor?
mlnmiles [1d\]: p1 a disater
masek [1d\]: 褉褜
bugtaker [2d\]: O1
Pu [5k?\]: not p1
gentil [6d\]: o4
masek [1d\]: hm
Sebass [7k\]: O-1?
blackrain [6k\]: w fine
Hagios [4k\]: o1?
Hagios [4k\]: Looks like the answer to me
TanGeng [2d?\]: t4
masek [1d\]: mistaked
]
;B[ns]BL[2965.008]C[gentil [6d\]: just o4
royalty [3d\]: nice
USGO15 [-\]: Ing Game, Round 4 will be starting soon. Live commentary from Guo Juan 5P
Squirrel [-\]: oups
Lothus [1d\]: lol
TwoStoned: white should undo...
Rakeliem [1d\]: is this a tougher mogo?
joelw [12k\]: o1
bugtaker [2d\]: lol
Lychee [7k\]: uh oh
TanGeng [2d?\]: wow
Himura [7k\]: o1 would sting now
Singer [10k\]: S2?
Olmerek [12k?\]: man, white is cheating ^^
SoniCJump [3k\]: owned
Amano [12k\]: white dead ; non ?
hakusai [6d\]: wow
kukulis4 [2d\]: oh no
TanGeng [2d?\]: ouch
Klever [10k\]: 锌芯锌邪 (
Sebass [7k\]: O-1 works!
Lothus [1d\]: nice bot!
shabam: o snap
Squirrel [-\]: it's not a tougher mogo, it's weaker white
gentil [6d\]: w cannot win this one
t0Strong4u [-\]: w is so behind
mlnmiles [1d\]: resign :D
blackrain [6k\]: ouch
Lever [12k\]: :(
chrono3450 [1d\]: this seems tougher because of the time - with more time, mogo can search more strings and find better moves
]
;W[rj]WL[3412.978]C[telchar [-\]: now this is like 15 handi :)
jhouse [3k\]: USGO15, what's MyungWan doing now? I assume not clapping
Singer [10k\]: Poor little pro :(
SoniCJump [3k\]: no match
royalty [3d\]: painful
Batavia [5k\]: maybe they shoudl try 4H next
Longshi [-\]: white wanted to test black more likely
prixdefer [1k\]: probably looking to leave it and start a ko or similar elsewhere, compter go no good at kos
chrono3450 [1d\]: W underestimated B
ketchup [6k\]: d6 overplay ftw!
silentfear [16k\]: gg no re
kukulis4 [2d\]: no just d6 was too early
ketchup [6k\]: then again, not unreasonable in handi game
Sakae [-\]: thinking machine 4 shows a good visual thingy for that kind of thing
kukulis4 [2d\]: but d6 still big
Kashuu [1k\]: white O3 overplay?
]
;B[gd]BL[2927.475]C[trumpet [3k\]: hehe, as i said. Mogo lured him into overconfidence ;)
Yxven [2k\]: silent uo?
Kashuu [1k\]: O2 works?
]
;W[gp]WL[3406.937]C[royalty [3d\]: mogo is stroong
zpt [2d?\]: trumpet, lol
Longshi [-\]: r1 was the overplay here
chrono3450 [1d\]: i told you, mogo is 2d level
change [-\]: how stong is this bot?
mlnmiles [1d\]: a 9p game in 'Active ga mes'
beer [1d\]: Mogo starts to behave like Deep Blue against Kasparov ^_^
Squirrel [-\]: mogo is not freaking 2d level
chrono3450 [1d\]: lol
Mocramis [9k?\]: bots ar
Mocramis [9k?\]: e
antipro: r1 was just a simply miss. but its funny that a pro can still make such mistakes
ibd [16k\]: too early to judge the bot imho
]
;B[gj]BL[2889.146]C[trumpet [3k\]: mogo will become insei soon
Tsuijin [1d\]: err
Longshi [-\]: maybe missclick
Yaxamie [14k\]: mogo could be 2d at 2 years / move
Amano [12k\]: beer=>chess, non ?
]
;W[dg]WL[3398.617]C[Zyphe [9k\]: science has proven many times over that the neurons in the brain are thousands of times faster and that processing on that scale is way higher then that of any computer.. one has to take into account what all the braind does.. it is crunching trillions of problems a sec just keeping the human body going...and thats before we try and think
TanGeng [2d?\]: s1 was correct right?
SoniCJump [3k\]: mogo 8 dan pro
antipro: r1 kind of shows white doesnt take the gmae really seriously
Zureiyaa [-\]: what did the pro just do in the lower right.........?????????
Mocramis [9k?\]: bots are invincible in litles tsumegos
Zureiyaa [-\]: is he really 1k?
sigmundur [4k\]: Zyphe: but the neurons aren't too dedicated on cruching go positions, on the most part
Zureiyaa [-\]: that's ridiculous...
Amano [12k\]: he's even 8p
Longshi [-\]: he missclicked I think
alexanro [7k\]: Zyphe: "science has proven" = show reference
Zureiyaa [-\]: can't watch this crap
]
;B[gg]BL[2849.858]C[ibd [16k\]: yeah zyphe, but keeping the body going doesn't help go :p
igojutsu [4k\]: nuerons are actually REALLY SLOW... but there are VERY MANY of them... that is why the brain can compute quicker
]
;W[cd]WL[3390.079]C[domn [14k?\]: Please stop with the armchair neuroscience/philosophy of mind and read some Daniel Dennet books or something do
Mocramis [9k?\]: you can't kill a bot in a without ko and really simple tsumego
chrono3450 [1d\]: daniel dennet can go learn some biology
TanGeng [2d?\]: neurons are associative so it's patterns that are easy
trumpet [3k\]: s1 inst of r1 was right for w imho
igojutsu [4k\]: <--- professional cog. sci. not armchair :P
TanGeng [2d?\]: lots and lots of connectivity
beer [1d\]: human thinking is not just about brain/neurons, there's something called consciousness which is not related to physical body
Keld [-\]: domn: Dennett is wrong, read Chalmers or Searle :)
]
;B[de]BL[2809.469]
;W[ce]WL[3387.862]C[Sakae [-\]: they are really fast but very few of them :D
chrono3450 [1d\]: beer - no proof of that
Sakae [-\]: new we have 3 theories
patrickb [-\]: keep commentary to the game itself, please
Sebass [7k\]: Oh god. Beer, stop now.
TwoStoned: yay patrickb!
Singer [10k\]: couldn't r1 have worked if he'd just connected at Q1 afterwards?
ekku [-\]: i think i need some beer.
tillydilly [-\]: lol @ beer
Articuno [-\]: Cmon Mogo, go super sayian and show him who's boss
igojutsu [4k\]: sure... this game looks pretty boring so far
CHUCHU [9k\]: how did he type in bold?
ibd [16k\]: beer kills my processors :p
TanGeng [2d?\]: q1?
beer [1d\]: Human is not just phusical, and that is why computers will never have intuition
Sakae [-\]: je
Dragonette [6k\]: no Singer, then black plays T2 and lives
Wizzos [4k\]: actually brain isn't faster. It just doesn't consider positions that seem obviously wrong. Can be tricky some times
]
;B[cc]BL[2768.385]C[Sakae [-\]: he's admin*
TanGeng [2d?\]: black lives in corner
]
;W[bc]WL[3385.63]C[PowerP [3d\]: w could still win
kuikentje [5k\]: this game is going to be a letdown =(
t0Strong4u [-\]: hard tho
w00t [2d\]: b feels the pain now
igojutsu [4k\]: hard to see how this game will get exciting since MoGo is not conscious ^^
Uberdude [2d\]: singer t2
Sebass [7k\]: beer.... Let's restrict our speech to facts :D
sigmundur [4k\]: don't feed the beer troll |:
progers [12k\]: where does B answer White's S2?
proutoz0re: beer you say that because you're not atheist
chrono3450 [1d\]: B is still way ahead, even in raw territory
Candidate [2k?\]: b dont feel pain..
Kosai [2k\]: progres: R1
Dragonette [6k\]: R1
Snowman8 [9k\]: what is this who programed mogotitan and made it so strong?
TwoStoned: i wish theb eer troll would feed me
]
;B[dc]BL[2726.437]C[trumpet [3k\]: r1 progers
]
;W[ch]WL[3377.949]C[progers [12k\]: ah, I see it now
Sakae [-\]: I don't see any beer..
beer [1d\]: every1 can prove it by using Kirlian camera
progers [12k\]: thanks
trumpet [3k\]: and then w cant approach
jtn [6k\]: http://www.lri.fr/~gelly/MoGo.htm
Sakae [-\]: oh there
Amano [12k\]: what do MS means ?
Stormer [2k\]: mogo was programmerd by a guy in china i belive
ibd [16k\]: snowman: mogotitan uses black magic, that's why it's so strong
Amano [12k\]: on their clock
Longshi [-\]: you don麓t need to be strong to program a strong bot
teytaud [-\]: mogo was programmed in France
igojutsu [4k\]: if it uses magic then it can be conscious
ibd [16k\]: yeah stormer, and I'm the pope
seku [1k\]: stormer: in france i believe
Amano [12k\]: france !
turingtest [3k?\]: knowing the common mogobot, this bot plays much more solid.
Stormer [2k\]: hmm
Sebass [7k\]: Whoa, you're the pope too?
]
;B[df]BL[2683.37]C[Stormer [2k\]: oh im sorry i am getting it mixed up with a different program
]
;W[cf]WL[3375.508]C[Dragonette [6k\]: Amano: do you use a non-english client?
Singer [10k\]: I thought I was the pope!
Sebass [7k\]: Maybe we're both the pope.
markboon [?\]: Whoa, the game started early?
TanGeng [2d?\]: have you played the common mogobot with 9 stones?
Amano [12k\]: dragonette)> what ?
ibd [16k\]: maybe Mogotitan is the chinese pope
prixdefer [1k\]: Anti-pope says watch game
Mocramis [9k?\]: the commom mogobot run on a common computer
LapinBlanc [2k?\]: Yep, mogo was programmed near Paris actually
Amano [12k\]: you mean, french ?
Dragonette [6k\]: I don't have MS on their clock anywhere on my client
sosuddenly [?\]: how strong is da bot?
Rakeliem [1d\]: c3?
turingtest [3k?\]: tangeng: hehe, no.
Dragonette [6k\]: I do have SD, which means "sudden death"
]
;B[eg]BL[2639.585]
;W[eh]WL[3373.751]C[Amano [12k\]: i use french client
Amano [12k\]: ah ok
Wizzos [4k\]: lapinblanc is the only other opponent of Mogotitan
chrono3450 [1d\]: W is playing really fast
markboon [?\]: Is this the pro playing?
t0Strong4u [-\]: r17 would b big
Amano [12k\]: "mort subite"
TanGeng [2d?\]: the motecarlo strategy changes with handi
Amano [12k\]: for MS
Singer [10k\]: I wish I was as good as either of these :(
fotland [?\]: yes, mark, pro playing.
markwebb [5k\]: white rabbit
TanGeng [2d?\]: of course this mogo bot will be better
markboon [?\]: I can't believe W R1
fotland [?\]: scheduled start was 2 hours ago
joelw [12k\]: f12
alexanro [7k\]: could we make MoGo play better by using several computers over the net?
Wizzos [4k\]: did you participate to its programming?
fotland [?\]: but they played a couple of blitz games first
markboon [?\]: where S1 is the standard way to win
kuikentje [5k\]: poor MyungWan
]
;B[fh]BL[2594.797]
;W[ei]WL[3370.472]C[Mocramis [9k?\]: at paris tournament (MOGO vs Catalin), Mogo lost with a enormous ko mistake
masek [1d\]: R1 was miss
Longshi [-\]: alexanro : internet connection would be too slow I think
TanGeng [2d?\]: yeah white can't afford this kind of exchange now
Singer [10k\]: would T3 have worked better?
alexanro [7k\]: Longshi: ok, thanks
Dragonette [6k\]: Alexanro: It'd be hard because of the latency involved.
Kosai [2k\]: are they playing one game only here?
Sakae [-\]: and packetloss
teytaud [-\]: Only one non-blitz game
Sakae [-\]: :D
]
;B[ej]BL[2549.622]C[Mocramis [9k?\]: does whithe panicked ?
lovebach [4d\]: why R1 i am trying to see why S1 doesn't work
jhouse [3k\]: they played 3 blitz prior, but only one long game
]
;W[fn]WL[3365.754]C[TanGeng [2d?\]: and you could say that black solidly owns about 1/4 of the board already
Candidate [2k?\]: I wonder if Mogo uses some algorithm to calculate where the randomised stones should be put, or if each position is as likely...
alexanro [7k\]: But, it's not much information being sent?
zeb [8k\]: oh man... you know what would be amazing?
zeb [8k\]: rengo with bots
Longshi [-\]: I think it does lovebach
turingtest [3k?\]: internet connection fails because of the sharing of the search tree dont work, i think.
Amano [12k\]: myungwan is very fast to play..waw
Longshi [-\]: but missclick is the best explanation I think
Singer [10k\]: could bots do rengo?
zeb [8k\]: why not?
TanGeng [2d?\]: bots could
Articuno [-\]: that'd be hilarious
hikaru78 [15k\]: i dunno
TanGeng [2d?\]: yeah
gandhi [-\]: i think it may have been better for Myung Wan not to play any rehearsal games with faster times against mogo
Amano [12k\]: singer==>could be fun
hikaru78 [15k\]: yea
trumpet [3k\]: teams with bot and human
]
;B[mf]BL[2504.331]C[gandhi [-\]: may-> might
Celila [-\]: candidate it's based on monte carlo
]
;W[mc]WL[3361.308]C[Singer [10k\]: wouldn't two different algorithms crash horribly to make it 25k rengo? :P
ibd [16k\]: lal
Articuno [-\]: exactly
Articuno [-\]: lol
Amano [12k\]: possible
TanGeng [2d?\]: nah singer no it won't
Mocramis [9k?\]: be fast look useless in absolute Timed game
Mocramis [9k?\]: it's possible
SpongeBob [5k\]: yeah - destroy black territory!!
TanGeng [2d?\]: i think the computer is basically memoryless
Candidate [2k?\]: Yes cecilia, but does it use some statistical method BEFORE the sampling?
hikaru78 [15k\]: go MYUNG
]
;B[nc]BL[2458.974]C[Candidate [2k?\]: say, 1/r from last stone
Mocramis [9k?\]: at the Paris' Tournament Mogo crashes two times
]
;W[nb]WL[3343.29]C[TanGeng [2d?\]: the monte carlo method isn't fully random
juhop [1d\]: it has patterns that affect random playouts etc
sigmundur [4k\]: Candidate: it has simple local pattern table
sigmundur [4k\]: read the pdf
prixdefer [1k\]: markov chains, blah just flip a coin
oloosh [?\]: mac play chess well
pasky [1k\]: my mogo things it is losing
alexanro [7k\]: turingtest: mogo is split in threads, so it should be possible to split it to several machines too, I would think.
pasky [1k\]: 43% probability of win
Singer [10k\]: If it was random, it would be 30k, unless there was a HUGE coincidence...
Candidate [2k?\]: tangeng, thanks, that was my question.
]
;B[lc]BL[2413.722]C[TanGeng [2d?\]: you have to account for the fact that games like this there are obvious single moves that win games
beer [1d\]: Is it possible to make a program which can LEARN go, starting from the rules and then learn 5-10 years, reaching pro-level after that?
Longshi [-\]: 30 kyu plays quite a bit worse than random
]
;W[oc]WL[3340.833]C[pasky [1k\]: moreover, my mogo would like to play r17 now
Wizzos [4k\]: could neurons nets be used to estimate te value of a move?
TanGeng [2d?\]: and those moves have to be played
lovebach [4d\]: Maybe againsr S1 (instead of R1), Black cut in S7
IanO: it integrates the result of thousands of random "playouts" to make a sort of evaluation function
Longshi [-\]: I mean, random doesn't push ladders that don't work
gandhi [-\]: wizzos, not very easily
McCrusty [1k\]: pasky - where did you get the program from?
pasky [1k\]: oh god, never mind me
gandhi [-\]: nobody has been able to do it
TwoStoned: please god, make the pseudo-science stop...
Dragonette [6k\]: Wizzos: runs into a sample size problem.
teytaud [-\]: mogo can be downloaded freely
pasky [1k\]: gogui is STILL not working properly :(
teytaud [-\]: (a limited version however)
alexanro [7k\]: beer: I think so. The challenge is to find a language to express the teching in.
]
;B[mb]BL[2368.488]C[Dragonette [6k\]: wizzos: and it's "nueral nets"
Singer [10k\]: I remember playing out ladders that didn't work, in the full knowledge that they didn't work
pasky [1k\]: I was trying to get that thing to work for the last 20 minutes :/
]
;W[od]WL[3338.105]C[Singer [10k\]: I don't remember why I did it
kuikentje [5k\]: there's a download section at the info page: http://www.lri.fr/~gelly/MoGo.htm
telchar [-\]: don't worry singer, lee chang ho makes that too
Articuno [-\]: driving tesuji!
]
;B[oe]BL[2323.167]C[alexanro [7k\]: MoGo rocks the score estimator ;)
]
;W[pe]WL[3335.314]C[Longshi [-\]: lee chang ho ?
pasky [1k\]: I gve up
Longshi [-\]: I remember Lee Sedol doing that
Articuno [-\]: duh, SE is mogo's little cousin
Longshi [-\]: but not Lee Changho
pasky [1k\]: I don't know how to ask my mogo about moves
telchar [-\]: ah yes it was sedol
Mocramis [9k?\]: of course, it's a handicaps game
Longshi [-\]: he knew why he was doing it though
]
;B[ob]BL[2277.904]C[Singer [10k\]: Lee sedol is cleverer than me, though ;)
Casillux [?\]: pasky: what is your UI?
blackrain [6k\]: lee sedol did the ladder and another pro did it before him
]
;W[pc]WL[3322.833]C[Longshi [-\]: kageyama did it too
Singer [10k\]: I can't help but feel that inspiration on Lee Sedol's level is wasted on go...
pasky [1k\]: casillux: gogui - got any hints on what else to use?
Casillux [?\]: gogui us good, let's chat in private
Yaxamie [14k\]: i figured it out
Yaxamie [14k\]: w did the right thing in the bottom right
]
;B[qd]BL[2232.599]C[seku [1k\]: singer: or anything else that is important to you =)
Sakae [-\]: huh Singer what do you mean?
lemurov [-\]: b'll lost on time
Nyanjilla [-\]: chat in private would be a good idea
Sakae [-\]: Lee Sedol isn't einstein...
seku [1k\]: :D
]
;W[qe]WL[3295.251]C[Pinchilei [2d\]: he is more clever
blackrain [6k\]: einstein wasn't einstein
Mocramis [9k?\]: Mogo become faster when the game is more closed
Yaxamie [14k\]: all of that unresolved aji in the bottom right is causing black to los lot's of processes rechecking the bottom
Yaxamie [14k\]: with every move
Stormer [2k\]: haha
Singer [10k\]: Perhaps the bottom was a devious computer-tesuji by w!
Mocramis [9k?\]: there are less options
markboon [?\]: Does B R17 work?
kuikentje [5k\]: haha Yaxamie
]
;B[qc]BL[2187.419]C[prixdefer [1k\]: br just ko material
]
;W[pb]WL[3292.171]C[alexanro [7k\]: Yaxamie: if you figured it out, where should white play? I think white died in the lower right
Longshi [-\]: that could be true Yaxami, except that black would have checked bottom anyway ^^
Pinchilei [2d\]: don't think it takes much time, not much combinaisons
Pinchilei [2d\]: also probably he doesn't recompute when no moves are played around
joelw [12k\]: einstein was actually jimmy neutron after a terrible experiment backfired and aged him 60 years
Stormer [2k\]: honestly if that were the case you think it would end up choosing another move there
Pinchilei [2d\]: I hope...
Stormer [2k\]: to remove the aji
Yaxamie [14k\]: nah, white loses it, but black loses maybe 5s per move
Yaxamie [14k\]: rechecking the corner
]
;B[na]BL[2142.144]C[EnHoeDan [4k?\]: what is the benefit of 8000 processors
]
;W[qb]WL[3288.485]C[patrickb [-\]: It
TanGeng [2d?\]: well the good thing the monte carlo method is that it's an easy parallel algorithm
markboon [?\]: Ah, W is lucky :)
sigmundur [4k\]: it's true that the "if A leads to B leads to C, play only C" is specially effective against a computer
Nyanjilla [-\]: please keep the comments focused on this particular game
Zyphe [9k\]: i wonder how well mogo is with ko threats
erice: more processors means extend more nodes
telchar [-\]: it scales well
patrickb [-\]: It's 800, and it lets the computer examine more random variations at each step.
Mocramis [9k?\]: Mogo could calculate more possibilities
TanGeng [2d?\]: so the 8000 processors are all doing simulations
Stormer [2k\]: what sort of machine is this running on?
Mocramis [9k?\]: and so check more moves
amadis [2k\]: last i heard mogo doesn't understand ko at all
Yaxamie [14k\]: it's running on LOTS of machines
]
;B[rb]BL[2096.743]
;W[re]WL[3282.17]C[amadis [2k\]: and wastes ko threats like they were nothing
kuikentje [5k\]: haven't you seen him do so last games, amadis?
Longshi [-\]: w messed up again ?
kuikentje [5k\]: he played zero point sente moves troughout the game
uurtamo [8k?\]: UL looks pretty big.
Stormer [2k\]: oh, they built a supercomputer?
ancientOne [2d?\]: considering the amount that is taking the computer to process this, really puts me in awe of the capacity of the human mind.
Mocramis [9k?\]: it's that who make it lost last time
Batavia [5k\]: so anyone counted?
joelw [12k\]: w is fine
Longshi [-\]: oh ok
]
;B[fi]BL[2051.414]C[telchar [-\]: the result of w invasion is :S
Mocramis [9k?\]: Human mind proceed by analogy
]
;W[ne]WL[3266.827]C[Singer [10k\]: C10?
ibd [16k\]: ancient: that's because go wasn't made for/by computers :p
yoyoma [1d\]: black R19 connect under ....
yoyoma [1d\]: *was* there
Longshi [-\]: ancientone : we use way better algorithm
BreatheGo [5k\]: how strong would this mogo be?
lovebach [4d\]: i don't think a computer losing time rechecking the bottom corner that is quite clear
Mocramis [9k?\]: and have a giant librairy of xperience
GoTal [10k\]: is it the best go robot?
lovebach [4d\]: we are soon in yose
ancientOne [2d?\]: it's just nice to be put in awe by something that i posess, u know. but anyway, back to the game
markboon [?\]: B S18 would have saved the corner
Mocramis [9k?\]: that mogo hav'nt
Sakae [-\]: d4 is still unbplayed
alexanro [7k\]: BreatheGo: mogo is as strong as the computing power you can throw at it
chrono3450 [1d\]: the bot is really strong in yose
chrono3450 [1d\]: so it can only go well for the bot
Articuno [-\]: I hope mogo doesn't throw away d4 haphazardly
]
;B[of]BL[2006.301]
;W[nf]WL[3263.513]C[Senkaku [7k\]: C11 better C10
joelw [12k\]: b has s18
uurtamo [8k?\]: p14 really bad.
HowToLose [5k\]: well, yose is a lot easier to calculate than the abstract middle-game
Mocramis [9k?\]: exepts in ko
Stormer [2k\]: actually cant b still connect under this way
kuikentje [5k\]: let alone opening, HowToLose
Singer [10k\]: How good would mogo be in the opening?
jhgf [5k\]: this game is very boring
Beanbag [3d\]: b need work on yose
Damir [3k\]: o16
master117 [6k\]: o16 looks nice
jhgf [5k\]: tho interesting in principle
]
;B[nd]BL[1960.955]
;W[og]WL[3257.121]C[Longshi [-\]: then don't watch it jhgf
memememe4 [?\]: no
Stormer [2k\]: probably not that good at the opening, but in a 9h game, no opening :)
jhgf [5k\]: cant help it
memememe4 [?\]: why not s17
zeb [8k\]: what just happened?
lovebach [4d\]: Mogo made a yose mistake in the 2nd blitz today, its not simple
arpad [1d\]: weird
yoyoma [1d\]: aha black can connect under and threat to save R17 next
Mocramis [9k?\]: R19 no ?
toxygen [?\]: pro used 5:43 minutes
toxygen [?\]: interesting
toxygen [?\]: he didn't take his time
Mocramis [9k?\]: it's sente
]
;B[qa]BL[1915.631]
;W[pf]WL[3254.5]C[zpt [2d?\]: everyone has experiended this lag?
Sakae [-\]: and you're 1d...
kuikentje [5k\]: I bet he's annoyed now that he ever accepted this challenge
tsiki [1d\]: meh, annoying when the pro doesn't take it seriously
Stormer [2k\]: lets see if it knows enough to torhow in at p15
TanGeng [2d?\]: so black can toss in at p15 now
TanGeng [2d?\]: ahh
TanGeng [2d?\]: nice
bagman [4d\]: i hope聽b plays throw-in
joelw [12k\]: why would he be annoyed now?
master117 [6k\]: q19 first
Sakae [-\]: he's starting to take it serous
Olmerek [12k?\]: toxygen, mogo thinks also when it is w's turn
joelw [12k\]: if he runs into some very good plays, he should appreciate them
]
;B[pa]BL[1870.309]C[pasky [1k\]: \\o/ I managed to work around all the gogui bugs
]
;W[cq]WL[3251.312]C[Stormer [2k\]: ahh shameful
Candidate [2k?\]: :O
Singer [10k\]: ?
telchar [-\]: no throw in
Mocramis [9k?\]: O_o
Sakae [-\]: now when he saw mogo knew he could tenuki
Yaxamie [14k\]: !!!!!
TanGeng [2d?\]: but it is a waste of points
Hagios [4k\]: o.O
giygas [4k\]: nice
Sakae [-\]: or he couldn't..
Mocramis [9k?\]: not so
memememe4 [?\]: what?
Sakae [-\]: but he did anyway
Olmerek [12k?\]: experiment ?
Stormer [2k\]: a 10kyu would get that move
Mocramis [9k?\]: he only lost the stones
Hagios [4k\]: Offering a trade?
pasky [1k\]: _now_ my mogo hopefully thinks it has 98% chance of win on the correct board
takayanagi [4d\]: w spent 6 minutes? lol
theremin [1k\]: C3 is enormous
Oshi [3k\]: 10 pts
memememe4 [?\]: I think he wants to cause the computer a malfunction
theremin [1k\]: and P15 10pts gote
progers [12k\]: uh!?
markboon [?\]: W is playing too fast
]
;B[oe]BL[1824.883]C[thefuture [5k?\]: 9 points
pasky [1k\]: H6 K6 K8 P15 S16 are best move candidates for my mogo
]
;W[iq]WL[3244.646]C[theremin [1k\]: around 10
royalty [3d\]: it's more than 10 pts
pasky [1k\]: looks like titan chose p15 instead, hmm
royalty [3d\]: about 17
Stormer [2k\]: more like 15
Yaxamie [14k\]: maybe it's a bigger move.
theremin [1k\]: but W gains a lot in the exchange
royalty [3d\]: 19 maybe
Pinchilei [2d\]: around 20
SadShape [1k\]: like 10 reverse sente
Sakae [-\]: more than 10
Stormer [2k\]: err ok ill defer to the 3d :)
Sakae [-\]: it's like.. 13....
joelw [12k\]: h6
bagman [4d\]: s11 is big
Singer [10k\]: Surely connecting was sente for w anyway?
ibd [16k\]: it's interesting that mogo comes up with josekis itself, without any databases about them :)
Sakae [-\]: no.. black lived
Mocramis [9k?\]: white can't win
TanGeng [2d?\]: hmmm
Singer [10k\]: Wouldn't be have connected at S17?
]
;B[jq]BL[1779.562]C[Schlaefer [?\]: no databases? sure?
]
;W[ip]WL[3239.522]C[t0Strong4u [-\]: k3 slowwww
TanGeng [2d?\]: hard for white to win now
Pinchilei [2d\]: it has a database
master117 [6k\]: mogo mogo^^
ibd [16k\]: no joseki database
ketchup [6k\]: s11 or k133?
telchar [-\]: no joseki database
progers [12k\]: Mogo doesn't have a database
sigmundur [4k\]: white still can win... black mistake, reduction in the top, big lower left
hikaru78 [15k\]: MOGO GO
ketchup [6k\]: k13*
Stormer [2k\]: b18 looking huge
ketchup [6k\]: ah wait c8 stlill there too
lovebach [4d\]: C3 is nice to see for amators i bet
Schlaefer [?\]: i would add a joseki database
alexanro [7k\]: wonder if mogo caches solutions that pop up regularly
toxygen [?\]: i doubt mogo doesn't know any joseki.
Kosai [2k\]: I'd play C18 rather than B18
Pinchilei [2d\]: he does have a precalculated positions database evaluated by himself
telchar [-\]: joseki databse weakens the bot :)
]
;B[jo]BL[1735.249]C[Sakae [-\]: no I think this is pretty much over
]
;W[bb]WL[3234.778]C[BigPhenix [3d\]: c6 black looks fine
pasky [1k\]: 0.984535(179%) || 62917/109552(57%)(40%)(51%/0.02) || 440 |18 || K5 N15 M14 S2 R1 C18 K2 L9 L8 M8 ||SSP:K5 J8 J5 E8 C10 ||PSP:K5 J8 J5 E8 C10
ibd [16k\]: yes, database makes it weaker because you can't follow them blindly
pasky [1k\]: this is snapshot of my mogo
zeb [8k\]: mogo learns joseki, becomes three stones weaker?
LapinBlanc [2k?\]: Well, it depends what kind of database, I implemented a big shape database that was learned automatically.
BigPhenix [3d\]: what does it mean, pasky ?
pasky [1k\]: 98% chance of win, k5 n15 m14 s2 r1... best sequence for both
seku [1k\]: mogo is learning robot, no? kinda natural then to learn joseki =)
Yaxamie [14k\]: ZEB: lol
Kosai [2k\]: ooh :)
pasky [1k\]: best moves for b k5 j8 j5 e8...
BigPhenix [3d\]: ok
Sakae [-\]: no
pasky [1k\]: that was before k5
Francisco [3d\]: thank you pasky
Sakae [-\]: monte carlo algorithim
TanGeng [2d?\]: best path algorithms require ways to evaluate every single board position
gandhi [-\]: mogo does not learn
LapinBlanc [2k?\]: Other people implemented manual shapes or rules databases in Mogo
kuikentje [5k\]: I don't think mogo learns, seku
]
;B[ir]BL[1692.063]
;W[hr]WL[3231.163]C[toxygen [?\]: pasky: well, these computations are basically useless.
kuikentje [5k\]: go ai isn't quite that advanced yet
seku [1k\]: ok
Singer [10k\]: Were the games against the 2k to teach it something about go before playing the pro, then?
og [9k\]: how has myung taken so little time?
Schlaefer [?\]: just thought because in chess it's all about opening libraries and such stuff
flex13 [9k\]: il y a du fight dans l'air
pasky [1k\]: toxygen: which computations?
Singer [10k\]: Myung saves time by losing groups
toxygen [?\]: pasky: they are done by the program which is struggling against pro
jhouse [3k\]: myung probably thinks on black's time :)
pasky [1k\]: toxygen: the program things it has 98% chance of win
pasky [1k\]: toxygen: do you think w is ahead on the board?
toxygen [?\]: well, good luck to him then
joelw [12k\]: b6 for b looks large
]
;B[jr]BL[1649.903]
;W[ef]WL[3228.069]C[toxygen [?\]: pasky: i never said w is ahead
lovebach [4d\]: Black is ahead
pasky [1k\]: my mogo suggested j5 instead :(
pasky [1k\]: toxygen: well you said "struggling"
BigPhenix [3d\]: does anyone know ? how many variations can this mogo calculate per minute ?
Sakae [-\]: well
telchar [-\]: e14 lol
Celila [-\]: your mogo is an older version
Sakae [-\]: he isn't too interested in this game
royalty [3d\]: 1700 000
BigPhenix [3d\]: wow
Mocramis [9k?\]: Players are generally weaker on computers no ?
toxygen [?\]: pasky: this means that it has to think very hard =\]
bearclaw: pasky: this statistic is based on the asumption that w plays randomly?
ibd [16k\]: 1700 000 per second.
patrickb [-\]: BigPhenix - it was doign 1,700,00/sec before
Swordfish7 [5d\]: how did white get so tricked in bottom corner
ibd [16k\]: not per minute
toxygen [?\]: mocramis: yes
Sakae [-\]: also he doesn't think his reading would help him in this situation
teytaud [-\]: pasky: our mogo has probability of winning against yours probably more than 98% (better software, better hardware)
]
;B[fg]BL[1608.91]C[pasky [1k\]: bearclaw: no, it is based on the assumption w follows the same algorithm
]
;W[ea]WL[3225.025]C[teytaud [-\]: thanks anyway for the numbers
trumpet [3k\]: w played r1 instead od s1
pasky [1k\]: teytaud: absolutely
AdrianG430 [12k\]: monkey jump!
jhouse [3k\]: my bot does 10,000 per second. 1.7M is nice
lovebach [4d\]: this game is now yose
giygas [4k\]: let's see how mogo will handle this
pasky [1k\]: teytauld: but they have similar ideas :)
SadShape [1k\]: how black will answer
SadShape [1k\]: ?
telchar [-\]: e14 has something to do with the monkey jump?
lovebach [4d\]: status is clear
tsiki [1d\]: my bot does like 100 =/
Stormer [2k\]: what hardware is this using?
Singer [10k\]: If mogo can calculate all those moves every second, how is it only 2k?
TanGeng [2d?\]: nah e1 is too far away
Candidate [2k?\]: swordfish: missclick/timesuji
Singer [10k\]: If I could read that far, I'd be a dan!
]
;B[eb]BL[1568.83]
;W[cl]WL[3221.716]C[Kosai [2k\]: Looks like B+50.
Sakae [-\]: its ineffective
toxygen [?\]: i can run mogo on my cluster and have it do the computations. can mogo run on irix?
ahnoeas [3k\]: it lacks intuition
egcdem:
LapinBlanc [2k?\]: The super-computer version is most probably better than 2k :)
flex13 [9k\]: c'st quoi cette merde
Stormer [2k\]: im still in shock that W could lose to a bot on 19x19
flex13 [9k\]: il glisse et puis plus rien
Kosai [2k\]: toxygen: They only published a Linux binary.
Celila [-\]: singer, the mogo playing on kgs regularly is not running on the same cluster as this one all the time ;)
theremin [1k\]: vocabulaire flex13
SadShape [1k\]: e19 was just to see how mogo will answer??
jhouse [3k\]: game is much closer than B+50
Kosai [2k\]: no source available.
lovebach [4d\]: pros are so good at yose
ZedGo [10k\]: flex13 -> vocabulary
Amano [12k\]: vive les francais
toxygen [?\]: kosai:64bit?
teytaud [-\]: toxygen: if you have a cluster and want to run mogo on it, you can email me (olivier.teytaud@inria.fr)
Francisco [3d\]: I have long prepared myself to see this
flex13 [9k\]: oui
Sakae [-\]: this wasn't a seruos game
bolb [-\]: flex13 : surveille ton vocabulaire
Francisco [3d\]: W will lose
mayday [1k\]: the bot cannot be 2k
]
;B[qm]BL[1529.779]C[flex13 [9k\]: ok
pasky [1k\]: my mogo wants to play d11
Longshi [-\]: flex : some respect toward the pro would be welcome
Sakae [-\]: because white missclicked
]
;W[rm]WL[3218.416]C[Kosai [2k\]: teytaud: Any idea on whether the source will be released?
toxygen [?\]: teytaud: ok, going to
mayday [1k\]: it's just not possible\\
mayday [1k\]: at the moment
flex13 [9k\]: desole
Sakae [-\]: white is just bored..
Kosai [2k\]: teytaud: It's very odd for a university to keep source code secret :)
Olmerek [12k?\]: I think white is experimenting
lovebach [4d\]: L13?
jnr [-\]: "If you have a cluster please call me " XD
markboon [?\]: S11 and big kill :)
]
;B[mh]BL[1491.674]
;W[nh]WL[3213.389]C[Mocramis [9k?\]: Myung could try to create ko, it's the last chance
pasky [1k\]: my mogo still likes j5 a lot
blackrain [6k\]: the sides are creating shapes to be exploited through sente push in yose
TanGeng [2d?\]: ha
Amano [12k\]: ^^
ibd [16k\]: my mogo thinks your mogo is stupid
pasky [1k\]: but would play k13 now
ZedGo [10k\]: are tesuji only for human being ?
]
;B[rk]BL[1454.603]C[change [-\]: t4!
]
;W[qk]WL[3205.83]C[lovebach [4d\]: is H6 better than J5 ?
pasky [1k\]: anyway, the probabily of win stays in 98%
Amano [12k\]: change==>霉ogo would see it ^^
theremin [1k\]: I wonder if B is not help by human from time to time :)
markboon [?\]: N12 was the right idea
pasky [1k\]: teytauld: does titan agree? :)
]
;B[qj]BL[1418.528]
;W[rl]WL[3201.894]C[LapinBlanc [2k?\]: You really have to know well mogo to know why he plays O11
czarny [-\]: good question pasky :D
markboon [?\]: but S9 was not
gandhi [-\]: going for the kill :)
Mocramis [9k?\]: but on what computer is your mogo running ?
gandhi [-\]: s11, then n13
joelw [12k\]: looks like just trying to get s11 in sente
Amano [12k\]: where can we download a mogo ?
TanGeng [2d?\]: yeah
]
;B[ri]BL[1383.141]
;W[sk]WL[3198.336]C[pasky [1k\]: mocramis: 2x intel xeon 2.4GHz
progers [12k\]: here it goes
Singer [10k\]: Is crazystone the same algorithm?
TanGeng [2d?\]: http://www.lri.fr/~gelly/MoGo_Download.htm
teytaud [-\]: CrazyStone and MoGo are very similar.
TanGeng [2d?\]: linux is the prefered os
SadShape [1k\]: idea is to kill p13?
Amano [12k\]: thxs
teytaud [-\]: (MoGo has a cluster version and CrazyStone not yet)
progers [12k\]: looks that way
Mocramis [9k?\]: it's tha same algorithm that the original mogo ?
TanGeng [2d?\]: ahh craystone too
Swordfish7 [5d\]: n13 ?!
]
;B[mg]BL[1348.783]C[hikaru78 [15k\]: gtg =(((((((((
]
;W[qh]WL[3195.7]C[royalty [3d\]: hehee
SadShape [1k\]: wow
puf [?\]: very nice...
royalty [3d\]: go for the kill, mogo
Swordfish7 [5d\]: but white is alive
Francisco [3d\]: MoGo is very good at killing
Francisco [3d\]: but it uses too much time
Kosai [2k\]: how is W alive?
joelw [12k\]: s12
trumpet [3k\]: t11 sente for ko for black
pasky [1k\]: my mogo wanna play l13
Rakeliem [1d\]: l13
pasky [1k\]: leisurely
joelw [12k\]: actually
ibd [16k\]: mogo will hopefully start playing faster
]
;B[qi]BL[1315.274]C[Swordfish7 [5d\]: t11 nothing
joelw [12k\]: s12 not good :(
Mocramis [9k?\]: bots are good in tsumego
Swordfish7 [5d\]: t4 is white sente
]
;W[kj]WL[3188.679]C[Mocramis [9k?\]: tsumegos
Francisco [3d\]: so maybe it is a disatvantagde
theremin [1k\]: surprising that mogo play this tesuji and connect N13 in order to kill ...
Francisco [3d\]: vantage*
GlebD [2d\]: mogo gona win !!!! ^^
cormorant [13k?\]: its funny that a computer that has time as a benefit... takes so much more time than human
Singer [10k\]: w+0.5 :P
trumpet [3k\]: oh right swordfish
theremin [1k\]: yes also
toxygen [?\]: teytaud: mail sent
]
;B[kk]BL[1282.572]
;W[rh]WL[3186.219]C[Batavia [5k\]: so is the mogo program aware of the time preasure it has?
ibd [16k\]: l11?
jhouse [3k\]: I wonder what a rematch with myungwan would look like. Would he still use only a few minutes for a complete game?
ekku [-\]: does mogo calculate on white's turn also?:)
joelw [12k\]: l11
pasky [1k\]: ekku: yes, usually
HowToLose [5k\]: this game is so confusing
lemurov [-\]: l11
pasky [1k\]: I assume they do have pondering enabled :)
]
;B[kh]BL[1250.678]
;W[lg]WL[3182.778]C[coolpapa [11k\]: /ponder on
SadShape [1k\]: white is alive?
giygas [4k\]: batavia, mogo is aware of time, yes
Stormer [2k\]: i doubt it helps much this game with w moving so fast
teytaud [-\]: pasky: yes, for sure :-)
jhouse [3k\]: T11 is life for white a t a minimum
trumpet [3k\]: well, now w will crush that moyo and then live
Sakae [-\]: SadShape yes
]
;B[li]BL[1219.712]C[Sakae [-\]: it's like 3 moves reading,,
]
;W[ki]WL[3179.969]C[chrono3450 [1d\]: i doubt mogo will be tricked by W's fast play xD
ibd [16k\]: what
joelw [12k\]: w not crushing any moyos
CoolHand [1d\]: computer should be able to play much faster as endgame approaches.
gandhi [-\]: there is probably a level after which more time does not make mogo that much better anymore
ibd [16k\]: pasky, still 98%?
teytaud [-\]: gandhi: not sure
gandhi [-\]: not sure, where it is for this setup
SadShape [1k\]: t11 is sente for black
theremin [1k\]: this kink of move is very surprising ...
pasky [1k\]: ibd: no
pasky [1k\]: ibd: 99% now
SadShape [1k\]: then i don't see how...
]
;B[ji]BL[1189.343]C[Stormer [2k\]: t11 t4
GlebD [2d\]: mogoooooo you rules !!! ^^
]
;W[li]WL[3168.99]C[toxygen [?\]: gandhi: with more complicated position the more and more "tricks" appear
memememe4 [?\]: is Mogo going for a kill?
Rakeliem [1d\]: well t4 is sente for white
Swordfish7 [5d\]: wasnt there a half mil dollar prize to whoever can make a program that can beat pro or something?
chrono3450 [1d\]: no
pasky [1k\]: ibd: ok, back to 98.7% :)
gandhi [-\]: teytaud, obviously it plays perfectly with infinite time and with very fast times it gets a lot better ith little time addings
chrono3450 [1d\]: i think mogo is trying to surround in a crafty way
Sakae [-\]: can't kill a fully living group
pasky [1k\]: my mogo wants m14
]
;B[lf]BL[1159.739]C[royalty [3d\]: t11 not sente against bottom
ibd [16k\]: thanks, pasky
]
;W[kg]WL[3166.371]C[Mocramis [9k?\]: yes
Stormer [2k\]: if b was really strong he would ignore t4 and kill w :D
pasky [1k\]: good that we agree
royalty [3d\]: mogo beats us
SadShape [1k\]: oh ok thanks rakeliem
Mocramis [9k?\]: but without handi ^^
ancientOne [2d?\]: Swordfish, i think that may be meant in an even game
Francisco [3d\]: without handicap
elgemelas [9k?\]: what a game
]
;B[jh]BL[1130.935]C[markboon [?\]: now t11?
]
;W[cb]WL[3160.715]C[tsiki [1d\]: Swordfish, it was $1.6mil but it was taken down in 2000
Swordfish7 [5d\]: oh well 500k dollars isnt worth much nowadays anyway
tsiki [1d\]: i mean the prize
royalty [3d\]: t11 tenukit10 t4
theremin [1k\]: W can't win now
HowToLose [5k\]: pasky: what are those percentages?
jhouse [3k\]: any strong players able to read how white would survive after B T11?
pasky [1k\]: howtolose: percentage of won playouts
ibd [16k\]: winning rate that mogo thinks it has
]
;B[si]BL[1102.979]C[HowToLose [5k\]: oh
Celila [-\]: there we go
master117 [6k\]: we will see ;)
patrickb [-\]: ow ow ow
HowToLose [5k\]: thx
master117 [6k\]: i think it kills
Pinguschaf [-\]: haha black doesnt get tricked
Singer [10k\]: T10, it seems to me...
Amano [12k\]: t10 , alive...
Sakae [-\]: ouch
Swordfish7 [5d\]: white is seki if not comletly alive
joelw [12k\]: w in trouble now
TanGeng [2d?\]: nah
Sakae [-\]: s8 is deing..
Swordfish7 [5d\]: no need for t10
SadShape [1k\]: t10 not needed
TanGeng [2d?\]: play the ko
Swordfish7 [5d\]: t4 is sente
telchar [-\]: b t10 dosn't kill
trumpet [3k\]: b may well sacrifice s1
Amano [12k\]: yeah too
royalty [3d\]: protip, white alive
Singer [10k\]: Failing that, T7 does the trick?
uuffy: nonetheless
Batavia [5k\]: pasky: how do you get those percentages. you have info on MogoTiTan?
Amano [12k\]: put yu loose one point with atari in >T10
Rakeliem [1d\]: t4
jhouse [3k\]: I was looking at the upper group
]
;W[rf]WL[3109.337]C[Stormer [2k\]: t12
Amano [12k\]: ^^
pasky [1k\]: batavia: no, I run my own mogo on a much weaker machine in parallel
ibd [16k\]: running his own copy and hooking up the game, I guess
shabam: t10 then 14
progers [12k\]: that's a professional ET
shabam: *t
Celila [-\]: it's not the percentages of mogotitan
joelw [12k\]: t10
]
;B[sj]BL[1075.539]C[pasky [1k\]: my mogo would play f18
Mocramis [9k?\]: but what was that C18 move ?
pasky [1k\]: but titan does not chicken out
shabam: isnt bottom dead
lemurov [-\]: t4
GoTal [10k\]: T4
]
;W[sp]WL[3097.456]C[HowToLose [5k\]: t4
zeb [8k\]: now we see if mogo understands ko
Amano [12k\]: he's dead...
SadShape [1k\]: no ko
Singer [10k\]: What does F18 do for anyone?
pasky [1k\]: it does very well
Francisco [3d\]: he really does
mayday [1k\]: no ko
Francisco [3d\]: it*
Principe [2d\]: is w doing it on prupose?
Damir [3k\]: 泻芯
lemurov [-\]: he's alive
Amano [12k\]: he had a dead group....
mayday [1k\]: T4 threaten bottom stones
lemurov [-\]: no ko
Sakae [-\]: isn't it ko?
]
;B[os]BL[1049.017]C[Senkaku [7k\]: alive
]
;W[rk]WL[3094.122]C[GoTal [10k\]: of course he live
Beanbag [3d\]: ko if w wants in sente
zeb [8k\]: oh wow
master117 [6k\]: s2 i s evil ^^
master117 [6k\]: was evil ^^
lukaszchol [13k?\]: ?
Principe [2d\]: wow storng robot
Swordfish7 [5d\]: maybe black losses on time
Sakae [-\]: ah yes
theremin [1k\]: alive gote
theremin [1k\]: game over :)
Francisco [3d\]: maybe
Francisco [3d\]: it麓s the only choice
ibd [16k\]: k13 next?
]
;B[fb]BL[1023.113]C[royalty [3d\]: f19..
PStillman [2k\]: there is no ko here
Francisco [3d\]: but it wont
]
;W[io]WL[3090.196]C[Singer [10k\]: w spends 14 stones on 4 points...
Wizzos [4k\]: no ko
Wizzos [4k\]: no ko id say
Noobita [16k\]: sword+fish=swordfish
Francisco [3d\]: to close to ooyose
bagman [4d\]: c11!!
Amano [12k\]: do computers accelerate their game and play bad when they ar in lack of time ?
Mocramis [9k?\]: why 18 ?
Francisco [3d\]: easy for mogobot
pasky [1k\]: amano: yes
Sakae [-\]: it's over
Mocramis [9k?\]: why F18 ?
]
;B[jn]BL[997.776]C[Rakeliem [1d\]: t4
pasky [1k\]: mocramis: safer ;)
ibd [16k\]: f18 probably safer
Francisco [3d\]: he is winning for sure
theremin [1k\]: now W will win on time
SadShape [1k\]: mogo is always ahead?
Mocramis [9k?\]: no
theremin [1k\]: of course
Francisco [3d\]: if it* plays like that
jhouse [3k\]: SE says B+45.5
theremin [1k\]: far ahead
giygas [4k\]: mogo will not lose on time
Mocramis [9k?\]: mogo will accelerate
Senkaku [7k\]: i think blk D12 before C11 is better
NattyIce [1k\]: K7!
]
;W[kf]WL[3054.42]C[Swordfish7 [5d\]: black win cuz white got so tricked at lower right
Singer [10k\]: Perhaps white would have fared better had he used more than 10 minutes of his time
Celila [-\]: mogo plays faster in the end
theremin [1k\]: oh ok
markboon [?\]: Big difference in level with the fast game
Amano [12k\]: pasky=>thanks ^^
Mocramis [9k?\]: if wh'ite moves are simple of course ^
Francisco [3d\]: I would strongly bet on MoGOTiTan any # of games
Sakae [-\]: it might very well have been a missclick
Francisco [3d\]: with 9h
]
;B[ke]BL[973.105]C[trumpet [3k\]: w trickd himself there
Stormer [2k\]: faster game favors the person giving handicap
]
;W[if]WL[3048.927]
;B[jf]BL[948.991]C[Amano [12k\]: tries to save E14 ^^ ?
makoregevr [2d\]: what is w lvl?
pasky [1k\]: 99.1% for b
Mocramis [9k?\]: does MyungWan used to play against bots ?
]
;W[jg]WL[3039.309]C[Suicide [8k\]: not necesarily true here, stormer
Celila [-\]: 8p
jhouse [3k\]: maybe this was fun for myungwan and he'll go out and buy a super computer to finally have a good computer opponent :)
Singer [10k\]: What percentage does mogo resign at?
teytaud [-\]: Pasky: thanks a lot, I feel better each time :-)
makoregevr [2d\]: b won
]
;B[ig]BL[925.528]
;W[je]WL[3037.401]C[markboon [?\]: Very good, B closes his territory
teytaud [-\]: Singer: probably 10% or 20% (I don't remember)
royalty [3d\]: too bad
NattyIce [1k\]: ko!
Beanbag [3d\]: its a ko
pasky [1k\]: teytaud: lol - you don't have access to the numbers?
Davzin: kokoroes
teytaud [-\]: no :-)
uuffy: no way w can win this ko though
pasky [1k\]: oh :)
teytaud [-\]: pasky: no :-)
]
;B[ie]BL[902.788]C[prixdefer [1k\]: br ko to osmal
Davzin: HIROGARU YAMI NO NAKA
pasky [1k\]: that's a pity :)
Swordfish7 [5d\]: no ko white just connect
]
;W[hg]WL[3034.468]C[Marek [5d\]: no ko
Davzin: KAWASHIATTE
Amano [12k\]: dubble
HowToLose [5k\]: m15???
Davzin: KAKUMEI NO CHIGIRI
ibd [16k\]: ZE WARUDO
Nyanjilla [-\]: Davzin---calm down
KGSPlus [-\]: please don't type in all caps
teytaud [-\]: I should have thought about that, I've forgotten - I don't want to stop the bot to add these outputs :-)
]
;B[ih]BL[880.471]C[Davzin: mkayz
Miles0000: WHY NOT?
pasky [1k\]: teytaud: hehe, understandably :)
Celila [-\]: ^^
Amano [12k\]: XD
Francisco [3d\]: The more he fights the more chance he has on time
Beanbag [3d\]: no ko
makoregevr [2d\]: why not really?
Beanbag [3d\]: lol
Marek [5d\]: to much atari
Singer [10k\]: mogo can't read the ko?
joelw [12k\]: decision time
Amano [12k\]: francisco=+>really ?:)
zeb [8k\]: m15! go for it!
Swordfish7 [5d\]: white is ok connecting
Swordfish7 [5d\]: he has no threats
Suicide [8k\]: how to lose...m15 k14 o11 n15, k?
]
;W[db]WL[2991.417]C[teytaud [-\]: it is often said that mogo is weak on "ko" situations
Francisco [3d\]: i mean, he has to fight
master117 [6k\]: k14?
bigbeam [4d?\]: huh
Francisco [3d\]: absurdly
Mocramis [9k?\]: could it work ?
trumpet [3k\]: s2 is a threat
]
;B[jf]BL[858.756]C[ekku [-\]: how much memory does mogo use?
Amano [12k\]: m15 is really godd
Davzin: This match is yummeh
Rakeliem [1d\]: threats g18
pasky [1k\]: hmm
uuffy: wth
Marek [5d\]: g18 nice now
]
;W[ni]WL[2979.802]C[Sakae [-\]: resign time
Swordfish7 [5d\]: ?!
theremin [1k\]: I guess W is tired to play vs a computer ... maybe he wants to go to have a break and sleep a little before his ext game
Lothus [1d\]: h 14... no ko
Suicide [8k\]: m15 is pointless
pasky [1k\]: casillux has a good point, i'm not sure if the komi is right
trumpet [3k\]: pls h14 now...
markboon [?\]: F19 now
Swordfish7 [5d\]: connecting was sente O-o
pasky [1k\]: damn :(
]
;B[hf]BL[837.51]C[Miles0000: Come on, MoGo can use the yoshimitsu!
]
;W[gb]WL[2977.111]C[SadShape [1k\]: ^^
Beanbag [3d\]: g18 wokrs?
markboon [?\]: W tricked B
royalty [3d\]: it's a trap!
Miles0000: no.
SunnyD [3d\]: lol
Miles0000: its a tarp.
Suicide [8k\]: g19
Amano [12k\]: ????
t0Strong4u [-\]: haha
Rakeliem [1d\]: what does yoshimitsu mean besides the sc guy
Davzin: Ask Sai for halp!
joelw [12k\]: h18?
Miles0000: i see what you did there
Swordfish7 [5d\]: black made 15 pts in the center off that
]
;B[fc]BL[816.846]C[Miles0000: yes
]
;W[fa]WL[2974.387]C[Amano [12k\]: coud someone explains me here ?
Swordfish7 [5d\]: i think he gains anyway
Miles0000: the soul caliber guy
prixdefer [1k\]: 8p bends the bullet
Miles0000: wait no
Davzin: Make that Akira child cry!
Miles0000: tekken
master117 [6k\]: h18
Rakeliem [1d\]: no, sc
flex13 [9k\]: noir a certainement perdu maintenant
Miles0000: no, its tekku
]
;B[gc]BL[796.772]C[Rakeliem [1d\]: , he was in 1 thru 4 probably
]
;W[hb]WL[2972.36]C[lovebach [4d\]: mais non
Singer [10k\]: Can w even things out with this invasion?
Miles0000: tekken*
chrono3450 [1d\]: :OO
Sakae [-\]: time for heavy reduction
Lothus [1d\]: lool
Rakeliem [1d\]: SC
Francisco [3d\]: no
Miles0000: no, look it up
Amano [12k\]: n'importe quoi
ibd [16k\]: j6?
RoryDix [3k?\]: go, myung wan
theremin [1k\]: B wins
master117 [6k\]: s + 28 Oo
uuffy: can w gain another 30 in yose
TanGeng [2d?\]: ... maybe he'll get it down to 10
Davzin: Yeah, 845-901-5005
]
;B[fl]BL[777.156]C[Miles0000: yoshimitsu is from tekken
flex13 [9k\]: il faut alors qu'il envahisse en bas a droite
theremin [1k\]: easy win
lovebach [4d\]: Blackstill wins
tsiki [1d\]: what are the percentages now?
okiol [6k\]: how strong is this computer?
Sakae [-\]: whites hopes just went up
]
;W[hm]WL[2969.928]C[kuikentje [5k\]: black is still ahead
okiol [6k\]: seems pretty strong
okiol [6k\]: xD
royalty [3d\]: what's the confidence of mogo now
Amano [12k\]: how many percentage, now ?
aburry [4k\]: when you're 8p, there are no bullets
okiol [6k\]: just arrived
telchar [-\]: now mogo humiliatingly allows reduction
bagman [4d\]: c11
Noobita [16k\]: makoregvr
pasky [1k\]: tsiki: I'm restarting gogui, checking if komi is right
Vogel [1k\]: why is b time running?
bagman [4d\]: c11
]
;B[hs]BL[757.931]C[markboon [?\]: B C11 very big
]
;W[gr]WL[2967.127]C[makoregevr [2d\]: yoshimitsu is the one with the lazer sword?
Noobita [16k\]: yes
Davzin: Um, go 845-901-5005, you can beat him
Beanbag [3d\]: mogo likely to lose
Miles0000: yes
makoregevr [2d\]: i like him -D
]
;B[in]BL[739.237]C[makoregevr [2d\]: =D
Singer [10k\]: pasky - what is percentage now?
]
;W[hn]WL[2964.265]C[Miles0000: hes freaking awesome
Swordfish7 [5d\]: i think good chance mogo lose on time
Amano [12k\]: pasky
Amano [12k\]: pasky
t0Strong4u [-\]: kik
Amano [12k\]: pasky
t0Strong4u [-\]: lol
Mocramis [9k?\]: i don't think mogo have the sense of humiliating
Francisco [3d\]: no
t0Strong4u [-\]: **
Francisco [3d\]: no chance
Suicide [8k\]: enough amano, pm him
Francisco [3d\]: i assure you
tsiki [1d\]: he's restarting the ui, just wait
okiol [6k\]: what ranking would mogo have at kgs?
Celila [-\]: don't flood, thank you
]
;B[ng]BL[721.096]C[Sakae [-\]: mogo will have 3 minutes left... or 5
Miles0000: but my favorite is nina
Amano [12k\]: xcuse
chrono3450 [1d\]: <_<
]
;W[il]WL[2955.872]C[jocelyn [2k\]: but it gained our respect, althought it don"t cares :)
Beanbag [3d\]: mogo likely to continue to make yose error
markboon [?\]: ???
Miles0000: because, well, you know
jhouse [3k\]: mogo has more time left than myungwan has used so far. It's easy to not lose on time in yose
TanGeng [2d?\]: ummm what was that move
theremin [1k\]: only chance for W is to win on time but some say it is impossible because mogo plays faster in the end ...
makoregevr [2d\]: nina looks good
Davzin: If I could...
Miles0000: i just unchied.
]
;B[sh]BL[703.174]C[Davzin: If I could play like that...
]
;W[sg]WL[2952.973]C[Suicide [8k\]: now if only mogo could win even...
Amano [12k\]: XD
egcdem:
Amano [12k\]:
Miles0000: Omg, jt would be pround
]
;B[cj]BL[685.909]C[Davzin: XD THIS IS UBER NOOB
ibd [16k\]: hmm this safer?
]
;W[bk]WL[2943.65]C[Mocramis [9k?\]: but while playing faster, mogo could make bigs errors
Miles0000: excuse me? more like boon.
Neospring [3k\]: give mogo 1hour each move and he'll beat myung o.0
Hinatachan [9k\]: what was the handicap set at?
GoGoDolls [7k\]: SE is so weird...
gandhi [-\]: b12
jhouse [3k\]: 9 handicap
pasky [1k\]: ok, if I *do* have the komi right now (I can't make entirely sure), mogo is still 99%
]
;B[dl]BL[669.143]C[markboon [?\]: I think the left was biggre than J8
Francisco [3d\]: no mistakes in yose
progers [12k\]: Hinatachan-9handi
]
;W[bj]WL[2938.749]C[Davzin: MyungWan is more azn, so he'll win
Hinatachan [9k\]: ty
Amano [12k\]: 99%
beer [1d\]: wasn't C11 better instead of C10... Mogo starts makin' mistakes
Miles0000: Davzin = epic boon
Suicide [8k\]: mocramis, I don't think the speed mogo is playing is affecting anything
Amano [12k\]: wpw
Amano [12k\]: wow*
Sakae [-\]: white might win
Francisco [3d\]: it has a very deep and fast yose calculus
Celila [-\]: it is suicide
chrono3450 [1d\]: B is making noob mistakes
makoregevr [2d\]: i like braien too
Yaxamie [14k\]: top right sacrifiece looks very good now
]
;B[dh]BL[652.691]C[Miles0000: No! Stinkchild, come on
makoregevr [2d\]: he have fast moves
Celila [-\]: just compare this game to the 3 blitz games before
jhouse [3k\]: nah, white has lost
]
;W[cg]WL[2932.74]C[bagman [4d\]: lol
ibd [16k\]: lol that move again
t0Strong4u [-\]: c11
Merlinor [9k?\]: white wins for shure ^^
Davzin: The more asian, the more likely to win
]
;B[jb]BL[636.749]
;W[ci]WL[2929.079]C[Amano [12k\]: XD
BattousaiX [-\]: *ahem*
Mocramis [9k?\]: of course, the more time mogo have, the better he is... if he have enouth memory
Miles0000: wait. did mogotitan make the word 'tit' like visible, for a reason
patrickb [-\]: Davzin - no more trolling please
shabam: is there a way to download and play mogo?
Stormer [2k\]: wow, what is this about mogo yose? totally missed double sente there
]
;B[di]BL[621.07]C[Amano [12k\]: what do trolling means ?
]
;W[jl]WL[2927.49]C[Singer [10k\]: Pasky - what is mogo percentage to win now?
Suicide [8k\]: shabam, google it
progers [12k\]: SE calling white top right dead
master117 [6k\]: jeah but i dont think you have 3000 processor or?
Almdudler [3d\]: some day, I麓ll be asian too
Miles0000: It means briging up random topics
ibd [16k\]: I bet it dropped a bit
]
;B[km]BL[605.821]
;W[ic]WL[2925.473]C[HowToLose [5k\]: se strong indeed
sigmundur [4k\]: mogo has at least used the time more efficiently...
markboon [?\]: I count 60 for W
IanO: mogo plays for safety, not points
TanGeng [2d?\]: are these Xeon processors?
juhop [1d\]: mogo does not care sbout points if it thinks it wins anyway
]
;B[ib]BL[590.967]
;W[hc]WL[2918.038]C[Miles0000: Stinko, get OUT OF THERE!
toxygen [?\]: tangen: i guess most likely yes
BattousaiX [-\]: SE+MoGoTiTan= strongest rengo team ever
uurtamo [8k?\]: they could be 8088's and it would still be good.
uuffy: it doesn't care about how much it wins by
toxygen [?\]: =\]
ww [1k\]: 9 handi is too much... seems like b wins
Sakae [-\]: hehe
]
;B[hd]BL[576.605]C[markboon [?\]: 70+ for B
]
;W[jc]WL[2915.187]C[toxygen [?\]: tangeng: it's said these details will be disclosed afterwards
alexanro [7k\]: score estimator was right ;)
WiseGuy [1d?\]: +30
Sakae [-\]: SE finds the weaks triple eyed groups and mogo kills it
Yaxamie [14k\]: huge reduction!
Mocramis [9k?\]: what is SE ?
theremin [1k\]: no more like +30
markboon [?\]: I think B will win
Stormer [2k\]: 9 handi and blunder in lowe right too much, i woudl bet on W in a rematch
t0Strong4u [-\]: l17 is sente
SunnyD [3d\]: mogo cares about winning margin
]
;B[kb]BL[562.483]
;W[gs]WL[2912.932]C[Singer [10k\]: Lee chanho and lee sedol would make an interesting rengo team
Suicide [8k\]: but Batt, SE would throw away live groups xD
theremin [1k\]: 25 maybe
Francisco [3d\]: i bet you Storneo
BattousaiX [-\]: xD
Francisco [3d\]: Stormer
]
;B[pl]BL[548.666]C[giygas [4k\]: i'd love to see black win by 0.5
Stormer [2k\]: ok :D
Amano [12k\]: i would say about 120 pts for black
Jenghis [2k\]: By my count black is going to win by 15 or so points
bagman [4d\]: b+ 10-13
Lothus [1d\]: lol
Francisco [3d\]: for real
Noobita [16k\]: opopopopopo
ww [1k\]: b might play moves in his territory when w passes so we dont know who wins
Dragonette [6k\]: I love MoGo's 1p gote moves
]
;W[ij]WL[2895.088]C[pasky [1k\]: ww: chinese counting
Stormer [2k\]: if they play again, and w loses, ill buy you a kgs plus subscription :D
ibd [16k\]: chinese rules anyway
Miles0000: if you blur your eyes, the white stones in the bttom right look like a, well, you know, wang.
]
;B[ik]BL[535.291]
;W[hk]WL[2893.034]C[ibd [16k\]: o___O
IanO: too bad the Ing prize expired. Looks like MoGo would have a chance at 9 stones
Francisco [3d\]: and what if w wins
]
;B[jk]BL[522.281]
;W[hj]WL[2891.453]C[Amano [12k\]: he becomes a god ^^
Suicide [8k\]: wow miles...wow
]
;B[rd]BL[509.606]C[PaperTiger [4k\]: ing prize wasn't for 9 stones
Stormer [2k\]: yes, i become king of kgs
pasky [1k\]: 99.87%
Stormer [2k\]: hmm
]
;W[se]WL[2886.89]C[Amano [12k\]: it's not 100.....
Stormer [2k\]: you must use a kgs picture of strawberry shortcake, for one month
]
;B[gk]BL[497.147]C[ibd [16k\]: lol
IanO: the Ing prize was graded; the last unclaimed prize was for 9 stones
pasky [1k\]: it almost never is 100 :)
Sakae [-\]: hmm
Amano [12k\]: lol
Singer [10k\]: 9 stones away from pro level must be KGS dan :P
]
;W[dm]WL[2875.966]C[ww [1k\]: so even if b played moves like o13, it still played better than 1k if he wins a 8p with 9h ?
lovebach [4d\]: Black is ahead 25 points
Sakae [-\]: white can't reduce anymore..
Sakae [-\]: white has lost
beer [1d\]: Next time, 9 handi for B + 29.5 komi for W would make really interesintg game
]
;B[gl]BL[485.131]
;W[hl]WL[2873.896]C[Zephon [2k\]: white can't reduce black anymore.. but black can reduce himself
progers [12k\]: remember, W misclicked
]
;B[hi]BL[473.332]C[ibd [16k\]: chinese rules
Mocramis [9k?\]: it's 100 when mogo could evaluate all possibilities
Sakae [-\]: no he can't
Sakae [-\]: chinese rules
]
;W[el]WL[2865.653]C[SadShape [1k\]: no if no white mistake in bottom right the game will be close
Yaxamie [14k\]: when did W misclick?
Singer [10k\]: Pro is unused to problems of internet play!
telchar [-\]: b played o13 because it was sure it could still win with o13, that's all
Amano [12k\]: chinese rules, what are they specific in ?
]
;B[me]BL[461.728]C[Sakae [-\]: doesn't matter if he plays in his own terretoru after dame
alexanro [7k\]: remember, misclicking or placing a stone in the wrong position is part of the game, rather
]
;W[of]WL[2861.992]C[jocelyn [2k\]: don"t forget white made some moves to test mogo's answers
Amano [12k\]: u can't take back your stone ?
Sakae [-\]: area counting not terretory
progers [12k\]: Yaxamie: W played R1 insead of R2
Yaxamie [14k\]: OH!
]
;B[od]BL[450.436]
;W[kl]WL[2857.756]C[Lever [12k\]: r1 instead of s1
progers [12k\]: yeah
Yaxamie [14k\]: i thougth it was computer tesuji
giygas [4k\]: how typical of humans, to come up with various excuses
progers [12k\]: S1 sorry
]
;B[ll]BL[439.589]C[master117 [6k\]: d9?
]
;W[dk]WL[2855.816]C[mayday [1k\]: if you plays bots like they are bots, they are easy to beat
Mocramis [9k?\]: he can't on a 9x9 so ...
]
;B[ek]BL[428.965]
;W[fm]WL[2853.542]C[Rakeliem [1d\]: ko
Celila [-\]: how often have you beat mogo yet, mayday?
Stormer [2k\]: the main differnece is chinese uses area scoring, you could the total area controlled by your stones and ignore captures... territory scoring is couting hte points inside and captures basically
TanGeng [2d?\]: at this point, a win is a win is a win so almost any move works
]
;B[sd]BL[418.626]C[giygas [4k\]: mayday, i think mogotitan would have a good chance of beating you in an even game though :)
kuikentje [5k\]: white's two weak groups are quite dissatisfactory ...
PaperTiger [4k\]: This is the last game?
proutoz0re: j1
]
;W[kc]WL[2846.493]C[mayday [1k\]: I never tried mogo, but I m pretty sure I could
gertzu [-\]: we want 9x9 match after this!! :P
Celila [-\]: yes papertiger
PaperTiger [4k\]: Pity
]
;B[kd]BL[408.528]C[Amano [12k\]: you do not count captured ? strange...
]
;W[is]WL[2842.058]C[Celila [-\]: well try it ... :)
Rakeliem [1d\]: Yea wgat happened to the 9x9
ibd [16k\]: 9x9 without handi?
Beanbag [3d\]: s1 actualy didnt work for w i think. cause b had q1 with that shape. b go s7 later to gain lib in semeai
sigmundur [4k\]: unless Myungwan wants a rematch (:
]
;B[js]BL[398.588]
;W[gm]WL[2837.275]C[sigmundur [4k\]: there might be room for one
Celila [-\]: it's only 19x19 today
mayday [1k\]: I should :)
sigmundur [4k\]: I think Myungwan simply underestimated mogo there
]
;B[qs]BL[388.889]C[Rakeliem [1d\]: myung won't need a rematch
]
;W[ck]WL[2833.877]C[sigmundur [4k\]: lower right for example, ouch
Elfwood [2k\]: white still hasnt used 15 mins total though
telchar [-\]: great achievement today
lovebach [4d\]: Humans are stronger at 37X37
]
;B[ia]BL[379.449]
;W[ga]WL[2831.566]C[Sakae [-\]: lol
Amano [12k\]: XD
markboon [?\]: yeah, w played too fast
Suicide [8k\]: amano, you count all the live stones and spaces on the board, so it's about the same, barring seki
Owy2001 [6k\]: Amano: You don't count the captures, but every stone you have on the board is worth a point
]
;B[ec]BL[370.236]C[Singer [10k\]: White needs to make game prolonging invasions :P
master117 [6k\]: mayb emogo plays als olng in his own territory until he loses ^^
]
;W[hs]WL[2828.996]C[Yaxamie [14k\]: lovebach: very true
mayday [1k\]: I used to play bots as part of a computer study ..
Rakeliem [1d\]: what the
weiqidevil [-\]: can mogo learn from games, that he played?
Sakae [-\]: yeah, but humans can fall asleep. which bots doesn't
Owy2001 [6k\]: So it's one point less, instead of one point more.
Rakeliem [1d\]: e17??
]
;B[ms]BL[361.405]C[ekku [-\]: what about damezumari in chinese counting?
]
;W[da]WL[2826.175]C[ww [1k\]: yes, these are the moves i was talking about :)
teytaud [-\]: weiqidevil: not efficiently
ZedGo [10k\]: loosing p2 made a big difference in the end ...
]
;B[ha]BL[352.776]C[TanGeng [2d?\]: dame counts
]
;W[ph]WL[2823.048]C[patrickb [-\]: K7/J7 worth points
TanGeng [2d?\]: I don't no about damezumari
kuikentje [5k\]: damezumari doesn't matter, because who gets it, gets a free dame point
]
;B[lm]BL[344.366]
;W[jm]WL[2820.846]C[Lothus [1d\]: lol
TanGeng [2d?\]: that doesn't apply
]
;B[kn]BL[336.027]C[ww [1k\]: -2 so far :)
okiol [6k\]: if he could someone would have invented AI imo
grolich [3k?\]: damezumari? black\\s groups has soo many eyes it can be considered a freak of nature:) unlikely
kuikentje [5k\]: *I mean, the ruleset doesn't matter for damezumari
Sakae [-\]: yeah right
]
;W[im]WL[2818.78]C[ww [1k\]: -3
Rakeliem [1d\]: why was e17 played
t0Strong4u [-\]: lol
Sakae [-\]: blacks endgame is superior you said..
master117 [6k\]: j11
chrono3450 [1d\]: 8p, and not even a normal 8p - a korean 8p loses a 9 handicap against a bot!
]
;B[gi]BL[327.941]
;W[ii]WL[2816.891]C[Lothus [1d\]: haha
Stormer [2k\]: the only tricky stuff is like, couting eyes in seki and strange things like that, i dont know those offhand
Amano [12k\]: not ure white losses
arpad [1d\]: not -3 but +0.001% chances of winning
Yaxamie [14k\]: so whats' the score
]
;B[mp]BL[319.897]
;W[ss]WL[2813.773]C[kuikentje [5k\]: it's endgame would be good if it mattered, probably
joelw [12k\]: wtf?
t0Strong4u [-\]: b wins either way even though he did -3 points
TanGeng [2d?\]: bot doesn't care
progers [12k\]: SE says B13.5
]
;B[hh]BL[312.277]C[Vogel [1k\]: yose is awful
aburry [4k\]: w gaining points here. why?
darkwings [6k\]: '讬讟 专 讗讬拽讟 讙诐谉诪注 讗讬谉讚?
master117 [6k\]: 13 points ^^ bad luck
]
;W[id]WL[2810.827]C[jhouse [3k\]: what's the difference between an 8p and a korean 8p?
Senkaku [7k\]: chinese rule
Rakeliem [1d\]: the bot is playing pointless moves
joelw [12k\]: n4 was the worst move I've ever seen
Makadoudou [5d\]: b +0.5
t0Strong4u [-\]: b making sure he won
Sakae [-\]: mogo is ahead
WiseGuy [1d?\]: Mogo usually goes for a .5 win
]
;B[ee]BL[304.896]C[hellboy [-\]: b+5
]
;W[rr]WL[2809.037]C[jhouse [3k\]: aburry - black won, so doesn't care
giygas [4k\]: people really should learn about the chinese rules
Lothus [1d\]: ....
Suicide [8k\]: koreans are revered for their skill
]
;B[ff]BL[297.569]C[Lothus [1d\]: disgusting
TanGeng [2d?\]: actually yose is bad because nothing it does effects its chance of winning
Sakae [-\]: it doesn't care if it wins with 361 or .5
Casillux [?\]: B only cares about victory not points
Elfwood [2k\]: which ruleset is most common in korea?
]
;W[]WL[2809.035]C[grolich [3k?\]: the bot is going to try to win by 0.5 or 1.5 as always
]
;B[]BL[297.567]TW[aa][ba][ca][ab][ac][ad][bd][ae][be][af][bf][qf][sf][ag][bg][qg][rg][ah][bh][ai][bi][aj][ak][al][bl][dl][sl][am][bm][cm][em][sm][an][bn][cn][en][gn][sn][ao][bo][co][do][eo][fo][go][ho][so][ap][bp][cp][dp][ep][fp][hp][aq][bq][dq][eq][fq][gq][hq][ar][br][cr][dr][er][fr][as][bs][cs][ds][es][fs]TB[ja][ka][la][ma][oa][ra][sa][lb][nb][pb][qb][sb][mc][oc][pc][rc][sc][ed][fd][ld][fe][ge][he][je][le][ef][gf][if][hg][eh][gh][ei][fj][fk][ml][nl][ol][mm][nm][om][pm][ln][mn][nn][ko][lo][mo][oo][po][kp][lp][np][kq][lq][nq][pq][kr][lr][nr][or][pr][ks][ls]C[joelw [12k\]: it's because mogo is on 5 min sudden death
takayanagi [4d\]: mogo always aims at 0.5 win
Yaxamie [14k\]: white is an expert of the reduction by filling dame
Sakae [-\]: it's like KGS system
markboon [?\]: B+10 + 12 handicap
Amano [12k\]: aaaaarf
kuikentje [5k\]: ouch.
theremin [1k\]: 1.5
HowToLose [5k\]: gg!
pasky [1k\]: fail
pasky [1k\]: it's not 0.5 :(
ibd [16k\]: gg!
t0Strong4u [-\]: hahahha
master117 [6k\]: 1,5?
Amano [12k\]: .....
progers [12k\]: gg!
Audwulf [16k\]: !!!
Yuura [8k\]: oO
Yaxamie [14k\]: !!!!!!
MaLa [2k\]: ROFL
Amano [12k\]: !!!
gertzu [-\]: yeah ;\]
NemoEimi [?\]: great! :)
Casillux [?\]: gg MoGo !!
sixsix [-\]: oh no!
Sakae [-\]: nice... althought the missclick must have felt hard..
Sagen [15k?\]: XD
telchar [-\]: great show
Amano [12k\]: great
mike2096 [5k\]: gg
mayday [1k\]: nice tho :) hehe
arpad [1d\]: gg
Yaxamie [14k\]: white would have won with no misclick!
patrickb [-\]: Thanks to the players
Yaxamie [14k\]: WHOW
Casillux [?\]: I am proud of you :)
trumpet [3k\]: BOTTIE RULES
Elfwood [2k\]: woot, the bots win
Mocramis [9k?\]: OMG
carwash [5k?\]: clap
juanra [15k\]: ooooooooohhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh
weiqidevil [-\]: pitty. but great game
Amano [12k\]: thanks
telchar [-\]: congrats to the mogo team !!
Tsuijin [1d\]: lets start mass producing super computer mogos!
GoGoDolls [7k\]: and SE still says B+14.5.....
change [-\]: thx!
Yaxamie [14k\]: *applause*
Celila [-\]: congrats MoGo-Team, thanks a lot MyungWan :)
giygas [4k\]: MyungWan, thank you for playing with mogo!
Stormer [2k\]: score one of the geeks :D
makoregevr [2d\]: bot bot =D
lukaszchol [13k?\]: I knew it!
kuikentje [5k\]: pretty nice
Beanbag [3d\]: u guys sure r1 was misclick?
pasky [1k\]: yes, mogo team is the best :)
kuikentje [5k\]: I'm impresed
mmueller [?\]: Congratulations!
pasky [1k\]: and thanks myungwan for nice game
Stormer [2k\]: for the geeks
rillig [2k\]: is there an explanation for R1?
dkiller [1k\]: thanks
kuikentje [5k\]: *impressed
IanO: thank you MyungWan
Nyanjilla [-\]: Thank you for an interesting game...
zeb [8k\]: wow
ZedGo [10k\]: thx MyungWan
alexanro [7k\]: remember that it was 9h :)
weiqidevil [-\]: Thank you MyungWan!!!!!
Senkaku [7k\]: Thanks for teaching myungwan
tsiki [1d\]: yes, what's up with R1?
Sakae [-\]: GoGoDolls SE is japanese
toxygen [?\]: pasky: that was 98%? :)
ekku [-\]: i'd like to see an even game
ibd [16k\]: Thanks MyungWan
markboon [?\]: Yes, very good
TanGeng [2d?\]: well the bot was definiitely sandbagging it at the end
sigmundur [4k\]: bots are ihQ
Stormer [2k\]: congrats
Francisco [3d\]: THank you MyungWan for the game
shoraishi [17k\]: bien jouer
Euphorie [-\]: r1 is surely a misclick S1 was the move
makoregevr [2d\]: Mogo Rules =D
TanGeng [2d?\]: it's probably better
sai38120 [15k\]: thankx a lot
pasky [1k\]: toxygen: huh?
Stormer [2k\]: thanks for the game myungwan
joelw [12k\]: fun game, ty
ww [1k\]: cant the mogo be programmed to play the yose more wisely ?
thoggi [1d\]: thanks MyungWan
Pinchilei [2d\]: Pros can let other players win by small margins to encourage them ;)
ekku [-\]: it would give much more info about real strength :)
toxygen [?\]: pasky: 1.5 points doesn't look like 98% to me =\]
joelw [12k\]: I'm sure it plays the yose well enough to win is its aim
TanGeng [2d?\]: you mean can't it actually count the points instead of trying to win?
Marek [5d\]: no it is...
Sakae [-\]: it was ahead
Beanbag [3d\]: i said that s1 dont work because of b go q1 and then go s7 in semeai to win
markboon [?\]: no, it's 100% now :)
pasky [1k\]: toxygen: I don't actually see your point - should it be more or less?
zeb [8k\]: it's a 98% certain 1.5 points
Singer [10k\]: Mog likes to win by a small margin ;)
ibd [16k\]: toxygen: it played safely because it knew it would win
Sakae [-\]: no need to play a better endgame
Merlinor [9k?\]: all because of the corner right down oO
kuikentje [5k\]: toxygen, it's 100% if there is no way to gain points for white
Celila [-\]: toxygen, as long as it was sure of winning, it didn't play moves to get points anymore
Marek [5d\]: win is win
ofir271 [6k\]: 1.5 is 100per
Euphorie [-\]: mogo is programmed to play the move with more winning chances, not the largest margin
TanGeng [2d?\]: Mog does random moves at the end if it's really far ahead
NPcomplete [5k\]: you have to admit that MyungWan didn't use very much time for this game. It was more like recreation for him! :D
TanGeng [2d?\]: yeah, the pro was definitely relaxing
chrono3450 [1d\]: a lax 8p is still dangerous
EnHoeDan [4k?\]: lets have an even against an 1 dan
lukaszchol [13k?\]: Was it historical moment?
ww [1k\]: it is supposedly good at LD, so filling own territory after all territory is taken ... ensuring victory is not a good explanation.. seen this happen many times
Stormer [2k\]: 8p could fight off an angry bear while continuing to play blitz
Yaxamie [14k\]: well Myung got to think durung Titan's long moves
Iceborg [6k\]: bot good
Beanbag [3d\]: s1 woulnt have worked for w. but maybe it would have helped w make eye shape faster and have a better position
Sakae [-\]: the bot was not intersthing thus blitzing
manestor [7k\]: i think R1 was to give mogo a chance, and it took it :(
fanertygr: 褌械锌械褉褜 卸械谢械蟹褟泻褍 褌芯褔薪芯 褋屑邪蟹邪谢懈 懈 锌芯褔懈褋褌懈谢懈
manestor [7k\]: like a 10th stone handi
tsiki [1d\]: last time MoGo lost because of a small mistake, now it won because of it =/
Steven1234 [5k\]: was top reduction preventable?
theremin [1k\]: thx MyungWan to play here !
change [-\]: what about 8p/ 6handi for mogo?
Sakae [-\]: yes isntead of taking the ko
Beanbag [3d\]: s1 is only working when w can aproach from the other side
ww [1k\]: probably 9h would be appropriate in a rematch also
beer [1d\]: but, chinese and japanese rules can differ at most by a point, and SE says B+14.5, minus 9 handi stones, gives 5.5, which is too much more than 1.5????
theremin [1k\]: no one thx the 8p to come ?
Sakae [-\]: mogo would be cruched
pasky [1k\]: theremin: plenty of people did
theremin [1k\]: oh ok did not see
Sakae [-\]: 9 stones is about correct.
SunnyD [3d\]: review?
Batavia [5k\]: very nice game
Beanbag [3d\]: maybe myungwan give review
juhop [1d\]: SE just counts wrong with area scoring as always
giygas [4k\]: let me thank MyungWan again for showing us these games
Sakae [-\]: but if the pro played seriously... maybe more stones..
ibd [16k\]: where is the misclick people keep talking about?
Stormer [2k\]: no, he should not review, we do not want mogo getting any stronger ^^;
Sakae [-\]: r1
ahmadr [10k\]: lol
Beanbag [3d\]: there no misclic they are rong
Batavia [5k\]: or maybe MyungWan could just give his opinion about the game without a full review
trumpet [3k\]: s1 q1 r1 t2 s7 works for w imho
Mocramis [9k?\]: reviews don' help mogo
Batavia [5k\]: did you enjoy this game?
NPcomplete [5k\]: his opinion would be interesting yes
Mocramis [9k?\]: mogo don't have Xperience
ahmadr [10k\]: Move 54
Stormer [2k\]: it might, i suspect mogo became self aware taht game
toxygen [?\]: i think iy myungwan played seriously, he would have taken more than few minutes to think
Sakae [-\]: its KGS sydrome
seku [1k\]: Sakae could be a strategic decision to play fast too... since mogo uses opponent time too
NPcomplete [5k\]: pull the plug Stormer ^^
Sakae [-\]: it's impossible to play seriously on KGS
Stormer [2k\]: :D
ahmadr [10k\]: MyungWan - is there any particular move that you played that you think set your game?
RoryDix [3k?\]: beanbag: do you have the sequence after S1?
Stormer [2k\]: i'm impressed though, i had no idea monte-carlo stuff could be so strong on a big board
EnHoeDan [4k?\]: lets play the comp 10 games even against 1dan
mayday [1k\]: if the bot is really 3k on a 19x19, it's pretty impressive !
ahmadr [10k\]: [move 54 = R1 - just click options->view offline and see for yourself\]
mayday [1k\]: as I m reading on their site
ibd [16k\]: stormer: this is heavily tuned/improved monte carlo
jnr [-\]: on a gagn茅
ibd [16k\]: thanks ahmadr
NemoEimi [?\]: Its little brother mogobot1 is already a stable 2k here on kgs, no?
jnr [-\]: on a gagn茅
jnr [-\]: XD
])