Thanks 

Got CCM title confirmed today. 

Since virtual impossible to get higher titles, despite lots of draws vs SIMs and IMs I can take more risk return to Dragon and try Kings Indian. As preparation I tried a unrated thematical KID event and to my very surprise engines still appears to have problems to understand these kinds of positions run in horizon effect in every game regardless of ply. I chatted with a IM in the event and he said that my oberservations are correkt and that computers always has trouble to understand these kind of positions and becouse of that he plays in rated games as well.

I can give soem deatails o how I play. I am an ameuter player. I have a CCE title but its the lowest title. But I am in a good position to get the next step, a CCM title, when 2 events are completed.

I can even tell what machine I uses. I actually only uses a laptop since I work from home there is no space for any larger extra computer. The laptop has a however a Ryzen 9 CPU and 32 GB ram, a gaming GPU card that I dont use, and a SSD drive where I have tablebases installed. This is also budget computer since the screen is smaller than on most gaming laptops. I imagine that professionals with with money and more space uses similar configuration but as stationary version = even more powerfull hardware. I use latest Stockfish as primary engine and in some positions I also check with latest Dragon by Komodo.

I played very solid Chess and in most position I uses depth 50 as minimum mark, except in some endgames where more than depth 100 and more is often reachable vert fast. There are some exceptions where some endgames actually gives lower depths but more accurate. I often get 50 ply very fast on my computer. There was some realy sharp games or some realy important games where I waited out depth 70 in a middle game position but I belive depth 50 is standard for most players but it depends on 
position and ambition. In the opening I use more time to check different book and bases suggestions a bit extra before deciding sub variations. 

EDIT The Fritz interface has a percent meusur on estiimsting max ply possible and in most positions says 100% at about ply 50-60 while in some positions depening on pierces left it can be much higher

Just curious,

But when you are running your engine of choice in CC tournaments, what is the longest you've ever let it run? Would you routinely let it chug along on a single position for 24 hours or more? If yes, how do you determine you've gone as deep as is worth it? This question could be for any who play CC, not just bragesjo. Thanks in advance.

[Edit: I realize that time & depth of analysis can be dependent on particular machine/engine specs. But I'd still like to hear your answers.]

bragesjo wrote on 01/12/23 at 10:45:40:

I was more referring to the fact that some players stop updating their engines/ databases and keep on using the old stuff. I am always surprised/ disappointed when I read about somebody using some outdated tools as that is just low hanging fruit.

trw wrote on 01/11/23 at 21:19:51:

For 2021 I have 2802 games with 51 decisive games.
For 2022 I have 476 games with 8 decisive games (my friend just sent me 2 dags ago December 2022 as it was not yet released at 8th of January).

This is purely related to the archive files from iccf which are published each month (which explains I just received December). Maybe you collect other games too (other correspondence organisation, other iccf files ...? Personally I find it strange you already have now 2602 games for 2022 while games still started in december and very likely aren't finished yet.

I used 2440 elo because that seemed to be the minimum in 2022 where each game is a draw (except clerical errors). You could debate about that figure as some decisive games are really poorly played so wondering if we can consider the rating as an accurate estimation of the playing-strength. Correspondence players at the end of their career often play hundreds of points below their old strength as they work only a fraction of the time anymore at their games compared with before.

I can give a little more of my thoughts. If a 2200 plays a 2600 in over the board games the 2600 is likely to win. But in Correspondence Chess I have seen 2000 players manage to draw 2600 particullary if lower rated has the white pieces.

When I had 2000 Elo I played 2 draws vs a back then top 6 rated GM in a unrated Petroff thematical event.

All NNUE network engines are already using self teaching AI in the evaluation that are trained in different ways. Stockfish, Dragon by Komodo etc are using this new technology