One of my favorite things about Tornado is the cool startup screen that flashes while you wait for the program to load. Tornado is named after the horse in The Legend of Zorro, and his programmer shows this cool looking screen at startup. One of the things I did not realize that was in Tornado I found out at Thanksgiving over my parents' house. I installed Tornado on my father's computer to test the new version. I don't have speakers on my computer, but my father's system does. Apparently whenever Tornado captures an enemy piece's knight, it lets a wild horse's whinnying sound rip through your speakers in an eerie echo. It reminded me more of Ichabod Crane's encounter with the Headless Horseman's steed from Sleepy Hollow.
It has a menu to turn this off, and I guarantee it would give anyone a jolt the first time they hear it!
This new version of Tornado got revenge over TSCP Gothic with a victory as Black. What was surprising was how TSCP "developed its Rook" to j3 after pushing a pawn from j4 to j5 to j6 on consecutive moves. I have noticed this beavior before from TSCP, and it is most unusual. Even more strange, TSCP was able to "predict" some moves by Black that I thought where unlikely, namely 17...Ab4, 18...Qa5 and 19...Ng6 back to back to back. (I just wasn't looking for a combination based on moves of this sort.) It amazes me that programs that are so different occasionaly agree on subsets of their principal variations, especially those that don't seem to be likely candidates to a human.
I will post the game later, as well as some pictures.
6 comments:
OK I know I'm just a dumb gurl and all but why did white play 34. K-Q1 and not 34. PxC? Couldn't white win the black Chancellor?
If 34. exf5?? black mates with Qg1+ Cf1 then Qxf1 mate.
This is going to sound weird. Even when smirf loses, it plays better than when most of the other programs win. I see so many wasted moves made by other programs (I'm not talking about Vortex, I mean programs other than smirf and vortex). At least smirf's play is "natural" and the way I think most of us wish we would play. I wouldn't wish myself to play the way some of these others programs do!
You know Victor I didn't have the words but you read my mind. The smirf program did play what I would call decent games. It's played much better than the other programs except its adversary. My bet is smirf will finish #2 clear ahead of the pack.
Fact is: Vortex is ahead. Where SMIRF would be sorted in still is open. I have seen a lot of structural weaknesses caused by its old initial basic ad hoc design, because of which I have concluded for long, that it is needing a complete rewrite. E.g. its evaluation routine might be realized then needing less than a third of its actual time consumption. Moreover cached values still are only reused unsufficiently. Neither 64 Bit processors' increased register pool is used, nor several processes on multicore CPUs.
Because of that it nearly has been wasted time to tune it up a last time for this event. It only would be justified, if SMIRF would manage to have some victories against those strong newcomers.
Until now I still have not begun to write a new SMIRF. Instead I did some considerations in its new basic design, and whether it should run on a Mac or a PC. Unfortunately there is very few feedback and nearly no support for to proceed with SMIRF as OCTOPUS then. Therefore I think, if anything it will be a very private and slowly evolving Mac project.
Maybe things could change having a solvent sponsor for Octopus, but none seems to be interested yet. Thus as an amateur I have not the chance to put that time into such a project as need be.
TSCP Gothic is a 64-bit program, as is Variant Pulverizer. No other Gothic Chess programs are 64-bit, including Gothic Vortex.
Porting to the 64-bit world will double your nodes/second. If your branching factor is greater than 2 (and whose isn't?) this means you will search 1 ply more in the same amount of time. That's all that you'll get for creating a 64-bit program.
No Gothic Chess program uses parallel processing at the present time. The playing field is, therefore, fairly level. And, all programs are being run on the same machine.
There is no hardware advantage for any program.
Programs will win or lose by their evaluation functions in this tournament.
Post a Comment