I played chess and then went to check what the computer thought.. As it turns out, I broke Stockfish Today, oops
The engine is inaccurate
Sort:
Can't even tell what is going on here since it doesn't look like you played rb1 in the game; you played nc2.
I do know you have a losing position after your fifth move as white, which is quite hard to do, so maybe you are not focusing on the right spot.
ArtNJ wrote:
Can't even tell what is going on here since it doesn't look like you played rb1 in the game; you played nc2.
I do know you have a losing position after your fifth move as white, which is quite hard to do, so maybe you are not focusing on the right spot.
That isn't the point, take a look at the evaluations, that's what is inaccurate, also, I wanted to see whether Rb1 or Nc2 was better after the game
If you didn't see it, the evaluation for the "Best" move is -0.17, yet the move that it says is "Inaccurate" is +0.29
Also, now just looking back, at 5. Nxd5, when checking, I found out that it is lost after 5. ...e6
Dude, the engine needs a little bit more depth. It's 10 now. It should be around 20-30 to analyze the position correctly. That's how things work in this life, man.
Are we talking about White's 22nd move in the game erubialgonzalez - ZestyZebra? In that game 22Bf1 and 22.Qh5 are much better moves than either 22.Rb1 or 22.Nc2
But ArtNJ is right, minor differences in stockfish evaluations are much less importance than the fact that you had a lost game with White after 5 moves! You might want to look into that!
Jan 24, 2020
0
#8
different engines will give different outputs...
here's mine -
1: erubielgonzalez - ZestyZebra 1-0, Live Chess Chess.com 2020
2kr4/pppq2rp/5p2/5p2/Pb2pP2/4P2Q/5RBP/R3N2K w - - 0 1
Analysis by Stockfish 230120 64 POPCNT:
22.Qh5 Kb8 23.Bh3 a6 24.Ng2 Bc3 25.Bxf5 Qd5 26.Rb1 c5 27.Rc2 Bb4 28.Bg4 Qxh5 29.Bxh5 Rgd7 30.Be2 Ka7 31.Nh4 Rd2 32.Rxd2 Rxd2 33.Bc4 Kb6 34.Nf5 Ka5 35.Ng3 b5 36.axb5 axb5 37.Bf1 Rd5 38.Nxe4 f5 39.Ng3 c4 40.Ne2 Ka4 41.Nd4 Bc3 42.Bg2 Bxd4 43.Bxd5 Bxe3
White is clearly better: +- (2.08 ++) Depth: 38/60 00:01:49 1987MN, tb=28039
(, 24.01.2020)
After 22.Nc2, it's nowhere near as good...
1: erubielgonzalez - ZestyZebra 1-0, Live Chess Chess.com 2020
2kr4/pppq2rp/5p2/5p2/Pb2pP2/4P2Q/2N2RBP/R6K b - - 0 1
Analysis by Stockfish 230120 64 POPCNT:
22...Bc3 23.Rb1 Kb8 24.Bf1 Qxa4 25.Qxf5 Qa2 26.Qb5 c6 27.Qb3 Qxb3 28.Rxb3 Rd2 29.Rxd2 Bxd2 30.Nd4 Rc7 31.Ne6 Re7 32.f5 a5 33.Bg2 Bb4 34.Bxe4 Rd7 35.Kg2 Rd2+ 36.Kf3 Rxh2 37.Nd4 Kc8 38.Rb1 Kb8 39.Kg3 Bd6+ 40.Kg4 Ka7 41.Ra1 Kb6 42.Rc1 Ka7 43.Kf3 Bb4 44.Ne6 Rd2 45.Nf4 a4 46.Rc4 c5 47.Nd5 b5 48.Rc1 a3 49.Nxf6 a2 50.Nd5 Ba3 51.Ra1 b4
The position is equal: = (0.22) Depth: 38/66 00:01:20 890MN, tb=67434
(, 24.01.2020)
Even the engine are often wrong. They're like really really good for the thing they memorize and for really really simple positions, but otherwise, they're not that good. Their main advantage over us and the reason they have beaten us on occasion is a psychological one rather than a logical one: we think they are invicible because they are machine and we lose contenance, wereas the machine do not fear us because it has no emotion.
I think why you broke Stockfish is simple here : you reached such a deep and rich position that iot got into a logic error loop.
1050 is already a really solid level, but you'll see, as you reach the expert level (1200-1400), it'll happen more and more.
I didn't even notice the depth 10. Stockfish gets to depth 20 in a small fraction of a second. Depth 10 is seriously hobbled and insufficient for anything other than spotting blunders.
This is true, but remember, I only get 1 depth 20 analysis every day I would say, And also, Yes, I analyzed that, and I learned from that, I hope I don't walk into a simple pawn push on move 5 anymore
WeakLava wrote:
As others have pointed out. Your time would be better spent not being busted by move 5, instead of a move difference of less then half a pawn.
Yes, I know, I shall avoid those now that I know
4xel wrote:
Even the engine are often wrong. They're like really really good for the thing they memorize and for really really simple positions, but otherwise, they're not that good. Their main advantage over us and the reason they have beaten us on occasion is a psychological one rather than a logical one: we think they are invicible because they are machine and we lose contenance, wereas the machine do not fear us because it has no emotion.
I think why you broke Stockfish is simple here : you reached such a deep and rich position that iot got into a logic error loop.
1050 is already a really solid level, but you'll see, as you reach the expert level (1200-1400), it'll happen more and more.
Engines don't memorize, they calculate...
Jan 28, 2020
0
#15
4xel wrote:
They're not that good. Their main advantage over us and the reason they have beaten us on occasion is a psychological one rather than a logical one.
I think why you broke Stockfish is simple here : you reached such a deep and rich position that iot got into a logic error loop.
1050 is already a really solid level, but you'll see, as you reach the expert level (1200-1400), it'll happen more and more.
So many false statements.
A. Modern engines are so much better than any human to ever live.
B. Computers are (mostly) better at analyzing complex positions than humans through raw calculation.
C. 1200-1400 is not expert level by any means. This is upper beginner-lower intermediate. 2000+ is expert.
Matt_Heup wrote:
4xel wrote:
They're not that good. Their main advantage over us and the reason they have beaten us on occasion is a psychological one rather than a logical one.
I think why you broke Stockfish is simple here : you reached such a deep and rich position that iot got into a logic error loop.
1050 is already a really solid level, but you'll see, as you reach the expert level (1200-1400), it'll happen more and more.
So many false statements.
A. Modern engines are so much better than any human to ever live.
B. Computers are (mostly) better at analyzing complex positions than humans through raw calculation.
C. 1200-1400 is not expert level by any means. This is upper beginner-lower intermediate. 2000+ is expert.
Expert on chess.com is 2400+