sje wrote:
First, I haven't gotten xboard to talk to the Citrine -- yet.
However, manual testing shows that the pieces on the board do not have to be moved when controlling the Citrine over the serial link. The obvious exception is when using referee mode.
Furthermore, the Citrine can be set to run in Auto or Demo mode where it plays against itself without the need for moving the pieces. The last game played can be retrieved over the interface. Well, at least the last 112 ply can be retrieved.
Sadly, the "S" (show running information) command implemented in some Novag models does not work. The Citrine LCD module shows this data, but it can't be accessed over the serial link.
None of the S1, S2, S3, or S4 commands work? Bummer.
Was the AutoNovag project I published last week of any help? It was sufficient to autoplay under WinBoard; it shouldn't be hard to adapt it to the more sane Unix-style serial devices.
I've also located the undocumented "J" command that tells the Citrine to start searching for a move. I don't think there are any more secrets to be discovered.
----
No response yet from FICS on getting an account for the Citrine.
----
The Citrine is actually kind of fun to play, even for a fish like me who has to use the easy levels. Not too much playing time is needed to become acclimated to the piece capture movement implementation. The auto sensory experience is to me far superior to square tapping, and I think this is more of a deciding factor than just having wood pieces or a "wooden" decal board.
Novag advertises the Citrine's strength at 2330 elo but its experiences with Symbolic show that Steve B's number of 2048 elo as being rather more accurate. Then again, the latter figure is measured versus computer opponents; perhaps a higher figure is called for when playing against human adversaries.
My Cookie Cat is quite unhappy with this new competitor for my attention. I keep the pieces safely bagged when not in use so they might escape her nocturnal predations.
ehh ¡ a lot of people said during a lot of moths that citrine have near 2100 points... and I... ( oliver ) was the first that "do the test BT-2650-2450 and do a extesive review...
the conclusion was. CITRINE have 2000~2025 points... after this... the firts evaluation a lot of optimist 2100, down to 2080... and now 2049... to do that ELO ratings use a rare games versus "other" computer with have a "sobrevalorate" elo ratings...
I say... before and NOW that citrine have 2000~2025... NO MORE.
IanO wrote:None of the S1, S2, S3, or S4 commands work? Bummer.
Was the AutoNovag project I published last week of any help? It was sufficient to autoplay under WinBoard; it shouldn't be hard to adapt it to the more sane Unix-style serial devices.
Yes, I tried all the Sn commands. I might retry some items without the Citrine LED module in line to see if that might make a difference.
My intermediate goal is to get the Citrine to talk with Lisp directly; success here would save a lot of time and effort in the long run. My hope is that this can be done without having to fiddle with system calls to set the line attributes.
Novag advertises the Citrine's strength at 2330 elo but its experiences with Symbolic show that Steve B's number of 2048 elo as being rather more accurate.
actually my number is the number reported by Selective Search Magazine in their most recent issue
Selective Search has been rating dedicated computers since 1985 and is highly respected by collectors worldwide
of course their main emphasis Today is rating Pc Engines http://www.elhchess.demon.co.uk/sscontnt.htm
it should also be remembered Steve ,that Novag's rating estimate of 2330 is a USCF estimate and not an ELO estimate so right there we need to deduct 125-150 points to get a comparable ELO figure
when Le Citrine first came out(2006) Selective Search showed the Citrine as low as 2025 ish or thereabouts.. but over the last year or so the rating seems to have settled into the mid 2000 level
will be very interesting to see how the rating settles in on ICC
will you be limiting your challenges on ICC to engines rated <2200 (or some other figure)?
if you let the top engines running on quads and octals play Le Citrine the rating will not be pretty
Steve B wrote:will be very interesting to see how the rating settles in on ICC
will you be limiting your challenges on ICC to engines rated <2200 (or some other figure)?
if you let the top engines running on quads and octals play Le Citrine the rating will not be pretty
Well, I still have to write the glue software and then get an account. FICS has not yet responded, and I'm not interested in paying for an ICC account just for the machine.
I will probably set the match formula to +/- 400 points from the machine's rating. This will let many humans match the Citrine; I won't leave the window wide open as that tends to attract 1200 elo humans that somehow know openings 30 ply deep and also play five man endgames perfectly. (Do they really think they're fooling anyone?)
The match formula will also have to be restricted so that the only time controls accepted are those that the Citrine can understand.
If I can get my old Novag Super VIP to talk to Lisp, I might also get that machine connected to an ICS.
sje wrote:
I won't leave the window wide open as that tends to attract 1200 elo humans that somehow know openings 30 ply deep and also play five man endgames perfectly. (Do they really think they're fooling anyone?)
it never ceases to amaze me how rampant the cheating is on all of the servers
i occasionally like to play unrated games on my Playchess account in the CAFE forum ...and even unregistered guests will cheat with engines in UNrated games
some how ..someway.. these folks get some form of satisfaction on seeing their temporary guest accounts win an UNRATED game with an engine instead of playing themselves
Steve B wrote:it never ceases to amaze me how rampant the cheating is on all of the servers
I don't care so much about the cheating as I care about those who cheat only when playing against programs. This screws up the program's rating and the errors can only be corrected over time by distributing the rating damage to honest opponents.