chessfags pls go
See what you did there
http://arstechnica.com/gadgets/2016/01/googles-ai-beats-go-champion-will-now-take-on-best-player-in-the-world/
They're both solved games with more pretentious tradition than demonstrable worth, so who gives a fuck?
>>45795254
>Go
>solved game
You're an idiot.
>>45795254
DnD will be solved next. The computers will start winning campaigns that weren't supposed to be won, or be played to be won. Just you wait.
>>45795293
Chess and go have a limited amount of boards. They're solvable.
AI's just good enough to play. That's the next best thing.
>>45794687
>>45795302
there's only a limited amount of campaigns and encounters. It's large, but limited.
>>45795254
Neither are solved, fuck you on about.
>>45795254
Even if they were solved, they have enough variables to the point where it doesn't matter for humans, and are actually competitive as there isn't any RNG. And there is nothing pretentious about either game, but there is in claiming that those games are beneath you.
>>45795254
Here's your (You).
>>45795336
Go is a 19x19 board. If you want to solve Go rigorously then please return after calculating all 361! board states. I'll give you 1.5*10^768 seconds to get back to me.
Solvable != solved.
Chess has more possible real game positions than go can have.
>>45796815
>1437923258884890654832362511499863354754907538644755876127282765299227795534389618856841908003141196071413794434890585968383968233304321607713808837056557879669192486182709780035899021100579450107333050792627771722750412268086775281368850575265418120435021506234663026434426736326270927646433025577722695595343233942204301825548143785112222186834487969871267194205609533306413935710635197200721473378733826980308535104317420365367377988721756551345004129106165050615449626558110282424142840662705458556231015637528928999248573883166476871652120015362189137337137682618614562954409007743375894907714439917299937133680728459000034496420337066440853337001284286412654394495050773954560000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
Age of the Universe
>432,000,000,000,000,000 seconds
I think you could do it if you could compute 1 state per plank time
>>45796782
>>45796815
>>45796886
>we /sci/ now
Please continue
>>45796815
>>45796886
Hold on, this doesn't account for pieces getting removed from the board and freeing up spaces.
Any ideas on how to calculate that?
>>45796935
Find a shape that takes exactly as many plays to fill as to capture and calculate an upper bound from there? But if you're doing something so optimized, you're obviously not taking every possible play.
>>45796935
It does account for that
I asked this a few days ago but got no answers; why does my tsumego app say that after 3, black is in the lead? Doesn't E9 kill black?
>>45796955
because at that point white can't stabilize territory on the wall?
>>45796951
no it doesn't, it treats every space as gone forever once chosen.
It also plays to every space, and I'm pretty sure you have less stones than that.
>>45796986
But white captures 7 stones and can expand left.
>>45796955
it does, but after that black can reverse the position by going D9
google Ko rules
>>45797019
Black can't D9 after E9 because of ko, and then black can't do anything to prevent white from playing F8. And there needs to be a really serious ko threat to avoid white taking F8.
This is the 10th problem in Tsumego Pro's 20k-10k Basic Tesuji set.
>>45797007
that just means it can't see through ko apparently
>>45796782
They have unfair terms from the start, this is mathematically shown. And I'm sorry, but both are used as a shorthand for "this person is a genius" and have nations staking their whole reputation on the players and centuries of symbolic reference, I struggle to think of ANYTHING AT ALL that is as fucking pretentious as those two shitty board games. Just because something is old doesn't mean it's worthwhile.
>>45796782
People think strategy is all about how simply moving units around and besting another player. Yet at its core, it isn't, its about dealing with uncertainty. Thats why the best strategy games have significant random elements, to create uncertainty... Most of the games people cite as "good strategy games" actually aren't strategy games at all. Chess is not a good "strategy game" neither is X-Com, starcraft or most RTS. They have their own strategy elements, but not games that require good understandings of probabilities and their effects on planning.
>>45797528
In go, white gets compensation points.
>>45795254
>solved games
I don't think you know what that means.
Neither chess nor go are solved.
>>45795336
To be fair, Chess is "solved" in the sense that a topflight chess AI can consistently beat the best player in the world.
From my understanding Go AI are considerably more primitive.
>>45799093
Actually they just made a Go AI using much more advanced techniques than the Chess AI and it now reigns supreme. It uses actual Artificial Intelligence rather than just brute force.
>>45796955
Black is in the lead, because that's only about a 20 point ko.
If it's the opening, black can certainly make more than 20 points with 2 stones in a row elsewhere.
If it's the midgame, then black can still make more than 20 points elsewhere.
If it's the endgame, then it's probably the biggest point value right now, but black *should* have ko threats worth more than 20 points.
So in any case, black can gain more than white. Remember to think globally and not locally. A ko is scary for new players, but if you use them right they win games by themselves.
>>45795293
The european champion of Go lost to a computer some time ago.
They actually didn't use bruteforce, but pattern recognition for it, so Go will be dominated by computers in a year or five. Probably two.
>>45799160
Ironically, last semester I took an AI course that explained what a bitch Go was to program and that's what I based my statement on. So I was right two months ago, and then Google came out of nowhere and smashed the European champ 5-0 when the best previous programs wouldn't have beaten anyone who got to the European tourney.
>>45799324
More like two weeks. Matches vs the world champion start the 9th of this month.
>>45799160
They made a Go AI that can sometimes win. It wins a lot less than you'd think, I believe it was something like 40-60%?
Still an amazing achievement, but definitely not "reigning supreme."
>>45799357
Oh, it was that fast?
Well then, Go will be solved game in two weeks, instead of two years.
>>45799357
We'll see if it actually wins this round. Wouldn't surprise me either way, especially since unlike Deep Blue, Google isn't likely to be a cheat to win.
>>45799345
To be fair, of 9 professional levels, google beat level 2.
Though to be fair, i don't know how strong level 2 to 9 really is.
>>45795708
It isn't limited. You could always add another goblin.
>>45799345
Fun fact: one of the first Nintendo video games was a Go video game.
>>45799160
AlphaGo beat Fan Hui 2p (2 professional dan levels) That's a huge achievement relative to previous Go AIs but a small achievement relative to the international Go world where top titleholders are 9p. Wait for the result of the game against Lee Sedol 9p, one of the world's current best players.
>>45795302
But they'll never beat our sworn enemy.That guy
>>45796815
Rekt
>>45800079
if anything they will be powergaming That Guys
>>45797665
Chess is a finite game, so it is solved by default.
>>45800128
But anon, after complaining for 3 hours about not being able to do what they want to do, they will choose tonot play the game
>>45800159
That's not what a solved game means.
>>45795336
You realize there is literally not enough physical entropy in the universe to solve either game (without quantum computer), right? Or are you shit with math? Most combinatorial games are unsolvable.
Number of valid chess games: 10^10^120, Number of valid go games: currently uncalculated, attempts to calculate have lead to a new classification of numbers (https://en.wikipedia.org/wiki/Surreal_number)
Number of valid go POSITIONS (not games): 208168199381979984699478633344862770286522453884530548425639456820927419612738015378525648451698519643907259916015628128546089888314427129715319317557736620397247064840935
Number of ways to shuffle a 52 card deck of playing cards: 8.0658x10^67
Age of the universe in milliseconds: 4.3x10^20
etc.
>>45794687
but this is our house
>>45795302
"Rocks fall, the CPU dies. Why? Because I am the GM, that's why."
>didn't see that coming, eh, mr robo-brain?
>>45801512
Error. My character immune from bludgeoning damage.
Action initiated; servo lift. Table flipping engaged.
>>45797528
No but it means it is sophisticated. No games have been more deeply analyzed than these two. This is where human strategies and tactics excel.
>>45797622
The uncertainty in these games comes from their complexity. You cannot truly calculate 15 moves ahead. But you can put your pieces into place increasing your likelihood of getting the upper hand on the enemy. You can suffocate an opponent through strategic movement in chess. Opponents who only play tactically by calculating 3 moves ahead will never understand why they lost.
>jesus, why does one have to explain such things here?
>>45797630
In chess players alternate colors.
>>45801551
No, I meant real world, not in-game.
>>45801787
>implying you can lift rocks
>implying you can drop sufficient rocks to break the CPU
>implying you'd wreck your friends/your own computer or something that's probably very expensive
>much implying many wow
