Best Up and Coming Chess Engine

[Werewolf]

Hi Jose,

yes that's right.
With many pieces on board the style is fantastic and very aggressive.

I will test your new version soon too.
Fizbo at first and your new version at second.

Thanks!

Best
Frank

I don't get your excitement.
It's a strong engine, quirky style of play, but that's it. There are others that are better at tactics and even though Hekkapelitta is improving fast, until it's in the top 5 there's just no knowing how serious it is.

Too early to call in my opinion....

[TShackel]

Hi Jose,

yes that's right.
With many pieces on board the style is fantastic and very aggressive.

I will test your new version soon too.
Fizbo at first and your new version at second.

Thanks!

Best
Frank

I don't get your excitement.
It's a strong engine, quirky style of play, but that's it. There are others that are better at tactics and even though Hekkapelitta is improving fast, until it's in the top 5 there's just no knowing how serious it is.

Too early to call in my opinion....

Hi Carl. I kind of agree with you. How "aggressive" an engine is ie how many wins it gets, just means it is stronger than the class of engines it is playing against. Put stockfish against Hakkapelitta and all its contemporaries and all of a sudden stockfish becomes the most aggressive engine with the most wins and the best at tactics.

I admire Frank's enthusiasm, and I share a similar enthusiasm for chess engine style. But to say Hakkapelitta is the best in tactics is just ludicrous. I don't agree with those assertions by frank. It loses against even the texel, hannibal, class of engines and even they are hundreds of elo points behind the top two.

What I really like to see is a chess engine play "smart". Somehow all its algoritions and evaluation function has made it stumble on to interesting plans in quiet positions, even without tactics. And when they play good in quiet positions and closed positions that's when they will really be lightning ahead of every aspect of human chess intelligence.

Sincerely,

Tim.

[TShackel]

Deep Fritz 15 !

Vas never renounced to develop Rybka ! I'm afraid ( ) that Stocky & Komodo must recuperate Elo points !
Just wait for a couple of weeks !

SilvianR

I really am excited for Fritz 15.0. I think Vas will be a better programmer for fritz than would horvath. I hope it's a long-term contract between the two. It'll be interesting to see what elo it comes in at for the rating lists. Thanks for your opinion.

Sincerely,

Tim.

[TShackel]

A private engine not to be underestimated is Ginkgo, which has appeared in this year's ICGA tournament.

http://chessprogramming.wikispaces.com/Ginkgo

Also, well-known regular Alfil has made very big progress recently.

Both engines participate in TCEC/Season 8 which starts on Friday. There, also a new Fire version 5b will compete.

http://www.chessdom.com/tcec-season-8-c ... formation/

So I think we have quite some candidates for the "best up and coming", and it will be exciting to see if any, and which one of them can really threaten the top-3

Yes, I saw that Ginkgo got very close to rybka strength. It's a shame it's not a public engine. But to come out of nowhere and suddenly be up there means it could likely be a clone. It's nice to see weak engines gradually over time work their way up the rating list. That shows originality and honest effort. Thanks for your ooinion though.

Sincerely,

Tim.

[Dann Corbit]

There are at least two acknowledged dotnet clones of stockfish that I know of:
https://github.com/bpfliegel/Portfish
https://github.com/mabicho/Stockfish-CSharp-Port

There are others with similarities as shown by ILSpy.

It would have been a lot easier to make a port of Gull.

Ports don't violate GPL, do they? ILSpy/DotPeek can indeed reveal the truth, more so if the "author" forgot to obfuscate the binary.

No.
In fact, they are open source projects themselves.

[Engin]

Hi Dan,

its not constant to make progress if you going the strength higher and heigher, at least it stick on the 2800 elo wall, to make >2800 elo any progress is much much harder.

[Dann Corbit]

Hi Dan,

its not constant to make progress if you going the strength higher and heigher, at least it stick on the 2800 elo wall, to make >2800 elo any progress is much much harder.

Yes, sure I know it. Elo is an exponential scale.
I am simply being over-enthusiastic.

It is really very difficult for any small team to overtake Stockfish, and the Komodo team has professional developers who can work 8 hours a day on the Komodo engine.

So realistically, it will be very hard indeed to pass either one.

But if an engine can do it, I think there are a few possibilities that have a realistic shot.

[Engin]

i guess that the top engines are not much different then 2800 engines, they are only very much tunned engines instead.

for example if they found for at least 1 elo progress for 1 parameter then they can get for 400 x 1 = 400 elo improvement.

but this is not easy too, even it cost many hours tuning to find any 1 elo gain.

[cdani]

i guess that the top engines are not much different then 2800 engines, they are only very much tunned engines instead.

for example if they found for at least 1 elo progress for 1 parameter then they can get for 400 x 1 = 400 elo improvement.

but this is not easy too, even it cost many hours tuning to find any 1 elo gain.

For Andscacs I can say that it arrived to more than 3000 thanks to having a lot or most of the know-to-work features. Tuning with three computers most of the day also helped of course.

[Dann Corbit]

Yes, if there is an Elo revolution it will come from search and not eval (like all of the other Elo revolutions).

Branching factor reductions are where the big leaps can come from.