|
|
| From | Message | ||
|---|---|---|---|
|
 question about computer analysis of our vote chess game game. I'm surprised the gameknot computer doesn't dislike this move more. It just gives a us a slightly greater advantage, going from a score of +1.24 at 45. ...Kg7 to +1.30 at 46. Qc5. THEN, the score moves to +2.98 with 46. ...Qf5+. What did the computer think we would play that it would not be more critical of 46. Qc5 ??? Is this not the losing move? Pete, other members, what do your analysis engines have to say about moves 45 and 46? Is this an error or weakness in gameknot's computer? Or is it normal? [got on to this from thinking: what if we played on from 45. ...Kg7 with Pete playing as white?]. |
||
|
bigpeta 20-Jul-12, 09:15 |
The oft forgotten horizon effect Up to 16 ply my deep Rybka 4 gives Qc5 as the next move. It is not until 17 ply that it switches to Qc1. From this i can assume that GK in common with many sites only has the engine check to ply 15. This is to save time and resources. This is the horizon effect. There are 3 important numbers to see in engine analysis and the ply depth. 10 ply will find blunders 15 ply will beat most players on any on line site 20 ply will give you the best move in 99% of the time. I have rarely seen an engine change its mind after 20 ply. So this is a lesson to be careful of engine analysis done by someone for you. Without the ply depth you have no way of knowing how accurate the results are. |
||
|
the-sigularity 20-Jul-12, 11:02 |
45 Qc5 of the end for white. My experience with computers is very limited (chess computers). But what I do know, is that these instruments are not always accurate. A computer is only as good as the data provided by the programmer. Thus. Garbage in, garbage out. I welcome critique. |
