97 ( +1 | -1 ) GK Forum v Fritz: Round 1I had this idea the other day and wondered if anyone would like to participate? Basically it would be the GK Forum v. Fritz 9, I've never played a long time control game against it, never mind a CC type game. So is anyone up for crushing the mighty Fritz? ...this is where I find out that everyone else beats Fritz in blitz everyday and it's just me who can't do it! ;)
I think maybe 3 or 4 candidate moves could be put forward with explanations, then everyone can vote on their preferred choice. I would then put the move into the computer and leave it analysing for a pre-set amount of time, then I'd post Fritz' move here. (I can set it so that I can't see what it believes our reply would be, that way I can still take part). It could be quite a good learning experience for players like me as we can openly discuss the game from the beginning - assuming Fritz hasn't advanced to the level of spying yet! ;)
Everyone of all levels are welcome, so any takers?
20 ( +1 | -1 ) Yes, I wanted to try this for quite some time. I vote for 1min/move for fritz,no tablebases. If that turns out to be too easy, we can increase the time - better than being crushed at the first try ;o)
90 ( +1 | -1 ) Good idea Matt ... just my thoughts on it...I like the idea of Candidate moves and voting for them ... as during the forum games earlier, it was simply too hard to sustain a concept thru numerous players of different styles and preferences. This way we could at least be sure various ideas are considered and can be more sustainable if chosen. Otherwise, I think it is just really hard to prosecute anything from a combination to a positional play without some meeting of the minds, in something so multiconceptual as a Chess position/game can be! }8-) PS.. like far1ey says too about mistakes, this way everyone had a chance to say "whoa" if needed ... and so maybe too someone does not have to be afraid of blundering if they want to put forward a really wild looking idea ... which are often the most interesting(!) If they hold up to scrutiny! So we can bring out the Tal & Velimirovic in everyone :))
104 ( +1 | -1 ) Let's start.Alright! Let's get this underway and see if things get up to speed. I propose we play as white to begin with, anyone of any rating can contribute and dont be afraid to make a suggestion since there's no such thing as mistakes, just learning opportunities (that's positive thinking for you!). I think it could work quite well if we begin by:
1. Proposing candidate moves (putting 'candidate move:....' or something like that in the subject), listing &/or explaining +'s and -'s (3 or 4 will probably do unless someone notices we have missed something better).
2. Comment and discuss existing candidate moves, giving our support or not.
3. Aim for a consensus on a move to play, if not, a voting majority will do.
4. I'll put the move into Fritz for however long we decide (put a suggested time in your post and I'll take a mean, median, mode or something along those lines)
OK, I'll take the easy one just to get the ball rolling and suggest 1.e5 as a candidate move: + Staking a claim on the centre + Opens diagonals for two pieces - ...? (I told you I was taking the easy one! :P)
23 ( +1 | -1 ) uhm 1.) e5!!That's a new one for me. But thinking about it - there definitely _are_ plusses. However, a BIG minus (which, you, at your rating are prone to overlook ;o)) ) - illegal move...
+ Staking a claim on the centre + Opens diagonals for two pieces
27 ( +1 | -1 ) Oops!If I ever meet the person that put 4 next to 5....! ;)
If I was going to start bending the rules maybe I should have suggested: 1. Qxe8#!?! + Wins the game instantly. - Your must get within punching distance of your opponent to play it.
But yes, 1.e4 sounds good to me. :)
4 ( +1 | -1 ) Sounds like fun!!I'm game!! 1. e4. Great start. : )
2 ( +1 | -1 ) 1. e4I bet Fritz playes 1...c5.
23 ( +1 | -1 ) 1...c5 is quite likely!Fritz has always played 1...c5 against 1.e4 whenever I've played it on full strength. How long shall we give Fritz on each move then? I'm thinking let's make it challenging and go for something like 30 mins }:D (...with tablebases?)
24 ( +1 | -1 ) Fritz starts tomorrrow :)If by tomorrow morning we have no more candidate moves with equal or more support than has been given for 1.e4 so far, then I will allow fritz to begin thinking. So let's show this heap of circuitry who is boss! }:)
24 ( +1 | -1 ) 1. e4!I was reading an archived article in an online magazine one time. The author made the comment about how Bobby Fischer would call 1. e4 "Best by test" and would sometimes even put an exclamation point on it!
30 minutes for Fritz, eh? Yeesh...
27 ( +1 | -1 ) NO!!!Don't underestimate Fritz! He plays at about 2700 in normal match play (=3min/move-average). I think we'll have a hard time to get even remotely near this level even with our best efforts. So I vote for 2min. MAX!!! And no tablebases.
57 ( +1 | -1 ) lol, ok! ;)lol, I guess you're right - 2 mins it is without tablebases. We can always try a longer time control if we fare better than expected. I'll post Fritz' move tomorrow morning (GMT), I'm off to my graduation ceremony tomorrow afternoon but I'll be back the day after, everyone should just go ahead as normal with the candidates for move two. If its clear what second move we want to play then if someone else has Fritz 9 (or less if they are the same strength) they are more than welcome to put it in under the same conditions so that we can get Fritz' second move.
29 ( +1 | -1 ) no one likes d4so i suggest a move on here?then you put it in the fritz?or something else.maybe i dont understand what too do, forgive me for being ignorant,then we talk about how fritz is going too kick our behinds,is this right?lol.example, we move e-4 he goes c-5,then we go knight f-3.i think.lol
63 ( +1 | -1 ) Fritz can be beaten inopen positions. But I'd think our chances better in something more closed, like 1.d4. Or feel good about playing a Ruy too. Computers generally don't seem to get much there. And the KI with a locked center is a good opening vs the box. Think we would do well in e4 if it is a Ruy, Scandinavian, French, Caro or Pirc. It is the Sicilian that worries me there, the most. 1.e4 probably will be the more interesting game though, so if that's what everyone likes ... as long as we Play Like Bobby, its best by test! :)) Shall we try following Bobby lines then, if it does play a Sicilian?! What do ya'll think about that?
67 ( +1 | -1 ) Why 1.e4 is best hereWe're not necessarily playing this game to prove we can beat fritz. We know that if we handicap it enough, out collective knowledge can win. So to me, we're playing this for fun, and we're even working out the handicaps to make the game as even as possible.
Playing for a closed position is a good way to beat a computer, but that's not our main goal. It's easier for everyone to understand the plans being made in an open position. We're more likely to agree if we can see what's going on. More people enjoy open positions and attacking and sacrificing. Yes, 1.d4 can be dynamic and the game can open up, but 1.e4 is more likely to be what we want and will enjoy.
Of course, 1.f4 is clearly the best opening move, but I'll support 1.e4 as well :)
6 ( +1 | -1 ) E4Just go with the flow... e4!
I was thinking about f3 followed by g4 but ... nah!
98 ( +1 | -1 ) QUESTION: Matt ...Does FRITZ have a 'randomizer' available ? For instance Novag used to have it in Constellation models. It still calculates best move, second best, etc. But then if the first two or three moves it likes are close enough in its strength evaluation, then it would randomly choose between the two, or three, as the case may be. Just thinking that might avoid any "repeat" of known wins against it. (Or will the Time/Strength control, or "style" if any, do as well for that? I've played Fritz, but am not very aware of all the features it offers.) *** ganstaman That's about what I was thinking too, about opening choice. The popular appear of e4 is without doubt and the play is the thing. While I feel d4 may be objectively better against a program, we will truly demoralize poor Fritzy if we beat him(her!?) at his best game!(I just hope that's not what Kasparov said to Deep Blue?!.. }8-))
51 ( +1 | -1 ) 1. e4 it is!Ccmcacollister I'm not sure if it has some kind of randomiser, I'll have a read through the instructions. I'm assuming it's set to max strength by default but I selected 'New>Position Setup up' and started with the default position, clicked on 'optimise strength' entered our move then set it on infinite analysis (it kept moving sooner than 2 mins if I set a fixed time!) and forced it to make a move after the right amount of time. I'll try the fixed time again next move & see if it uses all the 2 mins. But anyway, the info you are all waiting for...!
98 ( +1 | -1 ) 1....e5!Fritz plays 1...e5! Quite unexpected, I thought 1...c5 was a dead cert.
The position so far: 1. e4 e5
I'll get the ball rolling as usual and suggest the highly controversial 2.Nf3! + Begins clearing the kingside for castling + Stakes a further claim on the centre by providing control to two squares, one of which has blacks e5 pawn on, thus limiting the potential number of (sensible) replies. - We have seen it a million times before...as has Fritz.
So as usual, please add any +'s/-'s or argue against/elaborate upon any of them. Also suggest any candidate moves you may have. I was wondering if anyone would like to play an unrated game that would follow the moves of this one? It would mean we could post the board # as a link and we could quickly look at it which would keep things going smoothly once the position gets more complicated. If someone would be willing to do this then just send a challenge my way named 'GK Forum v. Fritz 9' or something like that.
11 ( +1 | -1 ) 2. Nf3 is goodAnd here is the bord board #5448348
26 ( +1 | -1 ) So we have possibly a petrov or philidor but most likely will have to choose between a ruy,italian,maybe scotch? I like the italian and some aggressive branches afterwards but fritz might eat that alive and make no opening mistakes.
28 ( +1 | -1 ) 2. f4Doesn't fritz use an opening book pretty much at random? Never know what bizarre defense it'll use against the king's gambit.
Plus, it'll be a sharp game where the forum members will have to be on their toes working on tactical nuances.
10 ( +1 | -1 ) I second mossberg. f4 would make for an interesting and tactical game. Let's beat the beast on its own turf.
3 ( +1 | -1 ) f4 Yes! Let's have some excitement! : ) : )
60 ( +1 | -1 ) Lets not forget ...2.d4 ... }8-) I'd recommend this path if looking for a Scotch/Goring type game. I'm torn about 2.Nf3 or 2.f4 . I'd like to play a Ruy. I'd Hate to play a Petroff. Think I'll abstain, so if it should happen, I wont bear responsibility for bringing another Petroff into the world ... On the other hand, if Good-Men stand by and do nothing; as in preventing a Petroff, evil can flourish ... and do I bear responsibility thru my inaction?! Possibly. On the other-other-hand, I already have 5 ulcers and can't change the Whole World ... but on-a-foot, this does fall within my sphere of influence ... so okay, I'll lean towards f4 ...
45 ( +1 | -1 ) Wow, Craig...When you mentioned "people saying bad things about the Petroff" I didn't think you were talking about yourself!!
From the white side, top players these days have been meeting the Petrov with 3. Nxe5 d6 4. Nf3 Nxe4 5. Nc3, Nimzo's old variation. Much different character of play than you find with the classical lines challenging black's N on e4.
If it's good enough for Topalov, what's your excuse?
50 ( +1 | -1 ) Oh, well in that case ...Jeff, you talked me into abstaining. While I'm not convinced anything good would come of Nc3 there, Nimzovich is dead after all, and whose to say it wasnt his opening that did him in, It looks like it could be fun enough. As for Topolov, he is the product of an over inflationary rating system, soon to be dethroned by Brunetti ! (so I havent bothered to challenge Topo :) But I am just THAT desperate to see a new (old) Petroff handling! }8-) Wait a minute, I forgot I DID challenge Topo already. (To a wrestling match, but he hasnt responded)
51 ( +1 | -1 ) 2. f4!I recently played a game with KGA and it was rather interesting, so I'm gonna go against my original candidate move (which was just to get the ball rolling really) and vote for 2. f4 now. Currently we have the following votes for the candidate moves:
I think that's right, correct me if I made any mistakes! I'll leave it until tomorrow morning again to give people an opportunity to change their mind or for new participants to add their vote.
7 ( +1 | -1 ) fine,never played it myself - but there's always a first time...
16 ( +1 | -1 ) I don't know..I foresee us being a pawn down on move 12 with no compensation and a tactical position that fritz thrives in. But it appears to be the majority so we may as well go f4.
18 ( +1 | -1 ) We don't have to play 2.f4, if you can convince others that it is not a good idea then we can still play something else - nothing is final yet! I'm open to suggestions still anyway. :)
9 ( +1 | -1 ) I don't mind either way. the KG in my opinion is a little objectively worse, but it may be more wild and fun as well.
37 ( +1 | -1 ) Should have played f4 on move 1 like I said...
Is the Vienna Gambit supposed to be better than the KG? If so, why not 2.Nc3 and wait a little to play f4? After all, "the beautiful thing about 2.Nc3 is that it threatens nothing" (Tartakower quote). We can follow up with either f4 immediately, or Bc4 (maybe get to play the Frankenstein-Dracula), Nge2, 0-0, and then f4. It may be a bit slower, but at least we're not gambitting our king :)
14 ( +1 | -1 ) I agree with sough.Lines such as the Tumbleweed are wild and exciting, leading to sharp play. It could be fun!
: ) Just how adventurous are we feeling in this game? ; )
41 ( +1 | -1 ) I'll give it a bit longer.Ganstaman has mentioned some possible alternatives to Nf3 and f4, so candidate moves are now: 2.Nc3 2.Nf3 2.f4 I have actually never played 2.Nc3 as far as I am aware so I wouldn't be able to offer any theoretically correct advice if we went down that route but I am happy to go with any of those three. (but I'm still leaning towards f4). I'll wait a little longer to see if anyone is interested in playing one of gangstamans suggested lines - leaving f4 to a later move.
13 ( +1 | -1 ) 2.Nf3I think that's the best move...a good, classical beginning. I've always felt that Nc3 puts black at an advantage.
266 ( +1 | -1 ) I like Vienna too ...Like ganstaman suggests, with a delayed f4. (I believe the actual Vienna Gambit would be 3.f4) . I am not very booked on it, but from my otb play with 3.f4 it always felt to me that I was needing a tempo for white; that a KGA type position from it, with N already on c3 was not good enough for me to play and win the opening battle. From Bronstein in his 200 Open Games book, I get the impression he considered the KG less of a risk than an actual Vienna Gambit, which he mentions briefly there on an occassion when he declined to go into it ... yet implies he would do so at another time. But I like the delayed f4 myself and my game with Dunne in 8th USCCC was such as that. (And the opening turned out to be very forgiving of WT's play, in that form! :) *** I like 2.Nc3 as well as anything in this position tho, except my own fav 2.d4 . If we want to go into a Bc4 opening tho, I do like the Bishops opening vs humans, since it has some tactical tricks ... that surely Fritz wont fall for, being a bit shallow but unexpected sometimes to humans. *** If we really want to win, I feel like Nf3 may be better than f4. My own thinking is that we can probably play f4, since Spassky played it vs Fischer. And Spassky-Bronstein is ... well if you never look at another Chess game, be sure to see that KGA first! My own bias to f4 is that it may give some opportunity to look at some really wild moves and sequences, which I personally enjoy. Not winning chances. *** Vs 2.f4 and KGA, GM Evans has suggested the Berlin Defense with ...Nf6 as being hard for WT to meet well enough to compensate for the pawn. While Fischer has suggested the way to go as being 2...exf4 3.Nf3 d6 instead of the 3...g5 he played vs Spassky. And then to play 4...g5 , believe it is. The var I always hated to meet was the Falkbeer CounterGambit , that I think comes out pretty even but WT hardly gets the kind of game he hoped for by 2.f4 *** If players wanted to post some games with KGA, that would surely be interesting and a good learning experience. Another positive point might be, that if Fritz really does stay in his opening book, rather than calculating, there is always the chance he will follow a line where an improvement for WT might be lurking. Or even a game unknown to him?! It would certainly seem to me to be a great opening choice for practicing ones analysis, with possibilities of some amazingly complex lines. (Most of which I will not know at all) *** Yet I am not lobbying for f4, and happy to see any reasonable move for #2. What is our Goal for the game ? *** [Maybe we should see how Fritz would answer each move and decide which looks most interesting?! ... just for one move? ]
143 ( +1 | -1 ) 2.f4 or 2.Nf3!"What is our Goal for the game ?"
Probably the right question. I thought we wanted to win this game... Or, more accurately, we wanted to test whether the specifics of CC (longer time control, use of books/databases) outweight the sheer calculation power of todays Engines. For that, we should try to play the best move every move. If we don't do that, we'll simply loose. Don't forget: we're playing against a SuperGM here. Would you play 2.f4 against Topalov? or 2.d4 against someone as accurate as Anand?
I guess the _only_ way to beat a computer nowadays is not to outcalculate it in tactical complications. We simply can't do that. For all our purposes, we can assume that Fritz will NEVER drop a pawn, overlook a fork, skewer or whatever. For the purpose of this game, we can safely throw our puzzle books out the window. The best we can hope for is to survive the middlegame and reach an ending where experience and intuition becomes a factor that will outweight calculation power.
Therefore I would strongly suggest that we limit our decision-making process to 2.f4 or 2.Nf3 Anything else may or may not be sound, but it has at least one drawback that shouldn't be underestimated: Far fewer games exist with those moves. There's a lot of experience here (and in books/databases) about what do do on the white side of a ruy/Petroff. And we can use every bit of it. There's probably considerably less experience here about what to do on the white side of a KGA. But how many experts do we have for 2.d4??
98 ( +1 | -1 ) Changed my mind! :PGood insights! I've been swayed back to Nf3 (I'm so fickle!), I would prefer to try to go for the win in this situation rather than my usual preference for sharp positions regardless of the most likely result, so all the remarks made make me feel that Nf3 is more sensible, if we can turn this into a strategic battle then we are more likely to be able to turn small positional advantages into one large enough to win as opposed to trying to out calculate a computer! I'm sure it will still be very interesting, but it will probably be more informative discussing strategic nuances than tactical variations - though I'm still not that sure... ;)
I think the current preferences stand at something like this: 2. Nf3 alberlie basbos sough luckypawn mattdw
It's still very close and I may have a couple of people under the wrong moves so I'll give everyone a chance to change their minds again.
162 ( +1 | -1 ) alberlie2.d4 is THE Move that Ive beaten Fritz with, in otb ... by out calculating it in a Goring Gambit. :) Or else it's book is flawed !? Because it errs in the very opening. *** Actually i might well play 2.d4 against a GM, as they would most certainly know the Lopez better than I can remember it for 45 moves. And I believe 2.d4 is a sound alternative, and a slightly better way for getting into Scotch associated openings than 2.Nf3. However, I would be inclined to sit out awhile if it were chosen here. Fritz seems to favor the same pawn grabbing loss line repeatedly. And we dont want a Guaranteed win, do we?! That's why I'd asked about a randomizer and have trouble considering Fritz to be a super GM. ** }8-) ** PS// I consider Fischer was better than Topolov. Don't you? Who can say ... but when Fischer was 2800ish, he was much further above his contemporaries than Topo. Do you think Topo better? I havent actually seen many of his games, as with Fischer. So my own opinion on that is primarily Elo & result based. *** mattdw ... about whether 2.Nf3 or 2.f4 ... Seems like alberlie and you have a good point there, to me. That the first seems the better winning chance to me too. But the second, I think we would probably draw and have a more interesting game and better learning experience, and better practice for combinative analysis. So I'll stay with f4 for now, if it is a choice there. Even tho we might need to work a little harder. If more votes come up for 2.Nc3 then I'd think about that possibility.
2 ( +1 | -1 ) f4 is very risky.I go for Nf3
1 ( +1 | -1 ) no further objections... (Craig, where are you???)So, let's go with 2. Nf3...
41 ( +1 | -1 ) 2. Nf3 it is!And Fritz plays....
...nothing yet! It appears to be broken, erm...I'll try to sort this out a.s.a.p - does anyone who has had any experience with Fritz know of a possible reason why the node rate has fallen from about 750Kn/s to just 15Kn/s?! It sounds like the hard disk is chugging (I dont think that's a word but it describes it pretty well), the whole computer becomes unresponsive and I just had to reset it. Hmm!
59 ( +1 | -1 ) just hit the "Fritz8" (or Fritz9) button in the engine window. A dialog pops up where you can change the hash-size. Just change it to something else (let's say 128MB) and hit ok. He should pick up speed again. Then you can change it back again.
The only explanation I can come up with is, that Fritz normaly uses the RAM-memory of the Computer. But there is virtual RAM also on your machine, namely that pagefile.sys in your root directory. THat is actually HD-space and accordingly slow. If Fritz is using that instead of the "real" Ram, it's likely to be _very_ slow...
26 ( +1 | -1 ) Alberlie's right...Decrease the hash size table or check to make sure you haven't got so many other applications up that you're using up all your "real" RAM. As alberlie says, once Fritz starts having to swap to disk for RAM, it gets extremely slow.
64 ( +1 | -1 ) mattdw: CHUGGING ?!Is that a "chunka-chunka-chunka-chunka" or . . . more of a "gur-eRR-gur-eRR-gur-RRR" ?! I'm pretty sure you dont mean MMM-mm-MMM-mm-MMM-mm or @swishh-@swishh-@-swishh . . . or guzzling pitchers of beer . . . :)) ******************************** alberlie Who; ME?? No objection here! I LIKE 2.Nf3 ...just dont want to be even partly Responsible if we get a Petroff and everyone goes "Uuuuugh-GoodBye" quicker than InstantKarma. As long as it is on your head if we have a Chess Apocalypse now, my cup runneth over. You did say we will Win with Nf3, yes? }B-) snicker-chuckle; chortling from the room . . . -> soundeffects