"It just isn't conceivable that you can design a program strong enough to beat players like me."

January 14, 2016

Stockfish 7 coming on strong to shake Rapidroid soon!

I just can't keep myself from sharing a snapshot of what is currently happening at the top of Rapidroid because the race between the two toppers appears not to be decided by a photo finish as it used to be since along time. There's a remarkable gap this time... 
Stockfish 7 should play even stronger than 121215 which climbed back to the top in the last Rapidroid version. It fulfills the expectation and brings some 50 more ELOs over the previous ranked one. Besides, additional help comes from Komodo 9.3 which plays unexpectedly weaker than 9.2. Consequently, the fish did not face any difficulty in dominating the field. The status is decisive after 7 rounds out of 17.

As soon as 17 rounds will be replayed, i'll be releasing the exact equivalent of the last ranking with the updated engines marked with blue in the image and add all the manual games stored in another pgn database.

It will probably shrink a little bit more as usual because of more samples taken into account. The error margins will get narrower but do not expect any radical change in the ranking.

Stockfish 7 doesn't give a chance to Komodo under Android. Sorry Mark and Larry...

Finally, it must be noted that ExChess 7.88, Crafty 25.0, RedQueen 1.1.98 and arm7 version of Phalanx perform all very well. I think they will all overtake their predecessors.


Anonymous said...

Yeah, SF7 on Android is much stronger than K, at least DF 160 version.
It dominated Komodo by a mile in my 40/4 mini tour, but now I'm running a mini tour with longer time controls (1800+2), as suggested by Alexei.
After 10/20 games (Nunn's test 1 openings), SF wins +6 -1 =3. I will finish the test maybe tomorrow, but I don't think the gap will change much :)

Unknown said...

Я думаю что один из недостатков komodo на андроид,небольшая скорость.Это проблема компилятора.

Unknown said...

Ok but Komodo 9.2 crushed all others including SF6 and lower nps than SF7 was the same. I feel that from 9.2 to 9.3 there are improvements which perform well with many cores on a PC but they disappear when compiled for Android. I think the issue is related to the code itself. It's made for extreme hardware...?