Komodo 10 - 60 ELO stronger?

Hi,

Komodo 10 will be available at May 30th and the programmer says, it will be 60 ELO stronger, than Komodo 9.

http://www.topschach.de/komodo-10-p-3483.html

Wow! :-) But what means 60 ELO for us?

A lot. It'll crush every engine 7/10

It means absolutely nothing.  All they have done is made the engine 60 rating points stronger than before, and its already unbeatable.  

incorrect, engine v engine battles are epic

I play official correspondence, and I prefer using Stockfish (git version) for the task. It's better/ faster at tactics, and quite good at positional play ( and in any case few strong correspondence players follow blindly the engine suggestions).

Apparently I'm not the only one- correspondence WC Leonardo Ljubicic has also expressed his preference for Stockfish.

Gonna have to disagree with you on that.

Also free :)

Quite certainly. Computers have outstanding combinational skills, but they do lack a few important things like long-term planning, sense of danger, and schematic thinking. Here is where the human factor pops in.

Maybe one day 95% of the correspondence games will end in a draw, but it won't be anytime soon... 

It means the programmers set it so that when there are 4 pawns vs 7, on odd numbered ranks, with a majority on dark squares, and the knights are facing sideways, on a weekday, but not before noon, the square e7 is valued 0.0000124 pawns more.

And after a billion practice games against the older version of itself, it scored a few % more.

lol

Been testing Komodo 10 out, it's not stronger than 9.42 or Stockfish 200516.. mostly draws.

Not worth buying - Stockfish is still the best as it's free. 

This rating list has Komodo 10 pulling away at the slower time limit of 40/40. 

http://www.computerchess.org.uk/ccrl/4040/rating_list_all.html

What time limit have you tested them at?

Not worth buying for whom? I'm pretty sure many chess professionals will want the best available engine for analysis. It's 60$ for Komodo 10, I guess you get all updates for free, until it reaches 11.

It seems Stockfish is better at faster time limits but Komodo is better at slower time limits. 

Doesn't that translate to meaning Komodo is better for analysis? 

The dirty secret is that modern 'correspondence players' have an engine look at multiple PV (like I do in chessbase for analyzing my openings) and go down the top lines a few moves...look at the engines subsequent 'evaluations' and...pick. If they want to pick something the engine thinks is not among the top lines, they use the engine to find out why.

In short, it's a crutch. A 'modern' "Correspondence GM", without his crutch could not compete with an average Correspondence GM from...say, 30 yrs ago head to head, without his crutch. A good engine and just 'good' chess knowledge/judgement, makes them all look like they are GM's.

Now as concerns Komodo 10. It's only a few 'elo' more than the old Komodo 9.42 which could be purchased before 10 was released. It's all about the annual TCEC tourney and trying to 'win' it as to when an 'official' new version is released. Heck, programmers can make changes between rounds. Its' all about winning that and as Stockfish is open source, the Komodo team ( I have owned multiple Komodo engines) have an advantage on the face of it.

What kind of analysis?

Stockfish seems to go into higher depth faster and is better for complicated positions/puzzles from my experience. Overall Komodo plays better  in longer time controls and gets a LOT less draws than Stockfish.

For modest CPUs and practical processing times, Stockfish is impressively strong.

I base my opinions on the TCEC results and my own tests with infinite analysis on the Nolot tests (note that 2-3 of the tests are unsolved and even refuted)

In 'infinite' mode analysis of late opening/early middlegame positions, wouldn't Komodo 10 be more accurate? 

Correspondence games use engines?