Houdini 5 has been released

[Modern Times]

The pricing model for Houdini is much better than for Shredder 13. Houdini standard version is for up to 6 cores, which satisfies most people, and then there is the higher cost version above that. Compare that to Shredder 13, where the lower cost version is 1 core only. That is a really outdated pricing model now.

[jhellis3]

I'm don't really think there should be two versions at all, but if so it should be on a physical CPU basis, regardless of core count.

6 is just as arbitrary as 1....

[Nay Lin Tun]

Current results of IPON so far! (235 out of 3300 played)

http://www.inwoba.de/H5.html

SF8- 3297
K10.2-3264
H5-3247
Far beyond stockfish, Oops.

But it is saying now
Houdini-SF 8- +10 -16

So SF8 is crushing it!

How do you know those details?

[rcmaddox]

What is the computer limit for the standard version? I have 3 computers and usually like to play computer versus computer.
Houdini 3 was limited to 3 has it changed?

The e-mail I received with my registration information says you get 4 activations.

[tpoppins]

Houdini 5 has been released today.
More information on our web page http://www.cruxis.com/chess/houdini.htm .
Also in the updated User's Manual: http://www.cruxis.com/chess/manual/index.html .

Make sure to refresh the browser cache using F5 or Ctrl-F5 to get the most recent page.

Thank you for reading,
Robert

Robert, did you manage to add back the greatest tactics that any chess engine has? Together with the great positional level, at the same time?

"Below some results for a popular tactical test suite with Houdini running at 20 seconds per position with 2 threads on a Core i7-5820K using 512 MB hash.

Arasan 19 Tactical Suite, comprising 200 medium to hard positions
- Houdini 4 (normal mode) solves 123 positions, total solution time 34:59 minutes.
- Houdini 4 Tactical Mode solves 158 positions, total solution time 24:31 minutes.
- Houdini 5 (normal mode) solves 124 positions, total solution time 35:30 minutes.
- Houdini 5 Tactical Mode solves 168 positions, total solution time 20:57 minutes."

dont know how these results compare to SF and komodo

On Kai's positional 200 Openings alpha suite H5 Pro performed on par with the leader, Komodo 10.1.

Congratulations to Robert on the new release and proving the naysayers wrong. I usually don't buy the Next Great Big Things the day of and wait a bit till the dust settles, but in this case waiting was simply out of the question. Three years was long enough! ;)

Break a leg in the Final! :D

[IWB]

[quote="IWB"
1. The input Elo is cmming from Ordo, the Classic is calculationg with simple Elo formula. That fits well with an average engine but it doesn't with one from the ends of my rating list.
2. It is better to compare average performance of the other engines. SF8 was at 81.5% and K10.2 at 78.3%. Both will lose some percent against H5. The percentage is the better indicator than the shown Elo.
...[/quote]

Just to demonstrate my point here:

After 2202 games (2/3) the Classic GUI with ORDO Data Input shows 78% with a rating of 3253 for H5. Behind K10.2
A propper Ordo calculation at this point looks like this:

Code: Select all

   2 Houdini 5                    :   3280     14   78.0%    37.1    3039      98        1717.5    1309     817      76    2202
   3 Komodo 10.2                  :   3264     10   76.5%    35.2    3038      88        3140.5    2418    1445     244    4107

So what is much more precise than the ELo is the %. I am thinking about not to give the "live" Elo calculation for future runs - or to use ELostat Elo for the live run as they would fit better in the Classic Elo calculation. (But that would confuse people to hve two different Elo values during and after a run ... ???)

Ingo

BTW: I really think that the contempt of K10.2 is a mistke. It should be 0 but it is 10 and that is what most people run their engine with.

[JJJ]

I think it's also a mistake. Now there is two engine at least where Komodo can't play with any contempt. And Komodo shouldn't probably play with contempt against Fire 5 and less contempt against Shredder 13.

It depend if the rating list is against top 10 engine of course or top 50 engines to use contempt.

[JJJ]

Houdini improved his score against Stockfish 8 and is now doing 45%. Very strong score and maybe can reach higher.

[IWB]

I think it's also a mistake. Now there is two engine at least where Komodo can't play with any contempt. And Komodo shouldn't probably play with contempt against Fire 5 and less contempt against Shredder 13.

It depend if the rating list is against top 10 engine of course or top 50 engines to use contempt.

I tested K10 with and without contempt against my TOP16 (Result at the start page, scroll down a bit). The "gain" or "loss" was minimal (same for SF).
At least against the TOP 16 it seems to be redundant and I doubt a big effect against a wider range as the lower 8 in my list already lose with a high percentage. Increasing the percentage for a winning engine is very difficult. If you play against even more engines with a wider range K10 will win 90%+ to add a contempt there to get 92% will gain nearly no ELO.

In short: If you apply a contempt to gain a very low Elo increase you can skip that as well, it will not change a lot but keep you "safe"(er) in case of playing against stronger engines.

Ingo

[IWB]

It aint over till its over

Ingo