Comments/Ratings for a Single Item

Capital idea! Demo games would be great!

Had it been a game I was already familiar with, that speed would have been fine.

Is there a way to adjust speed and pause play?

Anyway, I think your demo idea is a good one.

Some buttons to pause and scroll through the game might be a good idea. You
should realize, however, that it is only this fast because the full game is
already on the server. If the game is actually being played, the rate at
which the moves would be displayed is determined by how fast the engines
produce them. So in a 5-min blitz game (300 sec), moves would come every 8
sec or so, which is 8 times slower than what you see now. If people on the
average would still consider that too fast, we can play 10-min games.

The fast playback mode is really meant only to catch up with a game in
progress, when you first tune in. It seemed better to have the moves come
by once a second, so that the spectator can at least get a vague idea how
the current position originated, than to instantly flash the current
position. But the display must be a bit fast to catch up with reality.

Once someone is watching, he typically request displaying a new game as
soon as the previous one finished, and at that point there are no stored
moves to display. And it is a good thing that, as soon as people click the
page, they immediately see something happen. If they would have to wait on
average 15 sec for a move to be made, they might already have gone
elsewhere. A 'slow replay' button might be appreciated, though.

In regard to any archived games, a method similar to that used here, would be nice, if not hard to implement: http://www.chessgames.com/perl/chessgame?gid=1044044

Note: At that link we see a java player that allows the user to click forward or backward in a game. The game shown is Fischer vs. Addison (Cleveland Open, 1957) Fischer was only 14 years of age.

OK, I finally got to setting up a live demonstration match. It can be
followed at

http://80.100.28.169/gothic/falcon.html

Currently, you can watch a match of Falcon Chess there, between two
versions of Fairy-Max: one programmed to value a Falcon higher than a
Rook, the other programmed to value it lower.

Let me know if the link works for you.

George, let me know if you object to using Falcon Chess for this purpose.

Neat. Notice how often in the middle game Rook attacks Falcon, then Falcon
moves to offense attacking the same Rook, their endless more or less equal
interaction. Notice moving any Knight twice in the opening to the fifth
rank is thwarted by many available defenses, despite unprotected Bishop
Pawns the particular array. Notice these are Falcon moves planned,
foreseen or subverted, able to be blocked chiefly by only two Pawns sometimes, but of course
any two pieces too(sometimes). The ''sometimes'' is because even two intervening pieces/Pawns do not always secure the block against three-path Falcon. Pure leaping Falcon-Bison instead would be execrable mockery, of only average
interest about like Amazon(BNR) or strengthened Nightrider(NN) or
Squirrel(NAD).

Sorry the link went down: stupid auto-updates rebooted my computer
overnight... I restarted it now.

And yes, it is difficult to find a satisfactory array where all Pawns are
protected, because initially the Falcon can't protect anything. If one
sticks to conventional castling and quasi-symmetry (corner Rooks and
central KQ), the only possibility is RBFN. This is awkward, though, as it
gives a bad conflict between developing the Knight and opening a diagonal
for the Bishop. Plus the Bishops would look each other in the eye.

It does not seem too bad, though, to have unprotected Pawns in this
variant. The Falcons are not superstrong pieces, and also take time to
develop. This much unlike Carrera variants, where the (BN) and (RN) can
get into play and attack enemy Pawns on the first move, and are
super-dangerous pieces even in solo action. In Falcon Chess, by the time
the opponent can muster an attack on your Pawns, they are likely to have
already moved to a completely different position.

I would also be interested to have some feedback on the graphics design.
In WinBoard I used the pre-existing Lance symbol (a wildcard piece, for
which WinBoard accepts any move) to represent the Falcon. In the html
page, I have of course infinite freedom, (the board is simply a table of
gif files) and provided 2 alternative representations. But I must say I
still like the Lance symbol best: it sticks out most clearly from the
other symbols. Especially the bird-like symbol is difficult to spot. This
might change if I would depict the entire bird, rather than just is head.
I don't like that stylistically, though, as the Knight symbol also only
depicts the head (as is the WinBoad Elephant).

The Cobra symbol was inspired by the way the Falcon moves on a crowded
board. It does not stick out as clearly as the Lance, but can still be
spotted at a glance, due to its characteristic asymmetry. Problem of
course is that it is not really compatible with the name 'Falcon', and
that the C is already such an overloaded letter. The V is much less used,
but a Viper does not make such a nice picture. S for Snake is both an
available letter and compatible with the Cobra picture. But renaming the
piece is a big step.

A Lance also seems to have little bearing on a Falcon. Lance woud not be
such a bad name for the piece either, as its moves stick through openings
in the crowd to fairly large, but limited distance. Another interpretation
of the symbol, however, could be a feather. With a vary small change, it
could actually be made to look more like a feather, and it would stick out
similarly as it does now.

So I am inclined to stick to the Lance-like symbol, and say it represents
a feather. A more detailed symbol set could make this more explicit. (Note
that WinBoard_F does allow redefinition of piece ymbols, for thos not
satisfied with the pre-defined bitmaps. All you have to do is supply a
Chess font for WinBoard to render the pieces.)

Muller is aware of following, so for other readers> Not overstrong, Falcon
is always second in value to Queen only, though arguably Rook and Falcon
equalise the last moves endgames. With RNFBQKBFNR, the Knight cannot reach
the unprotected Pawn in three. Cheops' FRNBQKBNRF is recommended by
Abdul-Rahman Sibahi (his first choice RFNBQKBNFR with unprotected Falcon Pawn) and others. Without cornered
Rook, Pyramids' FBRNQKNRBF also protects all Pawns. Templars' RBFNQKNFBR oddly protects them all preserving corner symmetry. One of the Falcon-centralised arrays is Osiris' RNBQFFKBNR protecting all. In
Fairy-Max I like all the representations for play with no difficulty, but
especially Falcon head (Game Courier has lesser head) setting off as it
does Knight.

I was out for the weekend, but I let the match run on unsupervised. When I
returned I saw it was still going.

I noticed that after 213 games, the program that values Falcon below Rook
has taken the lead, 120-93 (56%). This starts to be significant: the
statistical error in 213 games is only ~3%. So this is a 2-sigma
deviation, meaning that it is highly unlikely (only ~2%) that it is better
to value Falcons above Rooks.

Ignoring PGN logs, we notice infrequent castling both g1 and h1. Over 600
games' 54% to 46% shows remarkable equality. Whoever first loses a R or F
indicates whether R>F or F>R per side maybe 80% of the time. Also Standings
tell whether it is odd or even game, as to which is F>R and R>F. In value,
Rook is to Falcon as Bishop is to Knight, apparently. Characterize N,B & R
as interacting. Likewise, N,B & F interact. Whereas opposites Falcon and
Rook contrast, rather than interact. Even same-side R&F, keeping their
contrasts, are hard imaginatively to try to get to ''interact.''
Orthogonal Rook's and multi-path Falcon's  different ways are at each
other's throat ever on opposite sides. The programs cautiously avoid
mid-term Falcon forks in fear of logic of intervening blocks, that fail to
materialize in real-world calculation by opposite number.  Also,
point-counting human player would tend to grab unprotected Pawn oftener,
using personal 1.1 or more for the Pawn.  [ Und deines Geistes hoechster Feuerflug Hat schon am Gleichnis, hat am Bild genug. --Goethe ]

I decided to terminate the Falcon-Chess match (I had let it run so long only because I was away anyway) after 676 games, at a score 356-320 (53%). Although this seems only a small difference, this is always to be expected in symmetric playtesting: they start out with equal material, and not all games will have R for F trades. And even for those that have: when I ran games that were imbalanced from the outset (RR vs FF), I never saw more extreme scores than 57% (and for only a single R-F imbalance, that would be halved).

Nevertheless, over this many games, a result of 53% is significantly
different from equality. One side scored 36 more wins than the other. The
standard deviation of the number of wins in 676 games is about 24. So the
deviation translates to 1.49 standard deviations, which has only a 6.7%
probability of occurring if the two engines were equal.

So the conclusion must be that it is almost certain (more than 93%
confidence) that it is better to value Falcons slightly lower than Rooks
than to value them above Rooks (in a simple scheme that has fixed piece
values throughout the game, like Fairy-Max). Current Fairy-Max has R=475, so I set F=450. (B=350, P=100.)

I have not had time to watch many of the games. (I will study the PGN
later.) Usually Fairy-Max is very materialistic: it has very little
positional evaluation, and the diffence between having a Knight on a1 and
on e4 is only about a quarter Pawn. If it lets undefended Pawns live, it
is usually because they are heavily poisoned.

This was interesting series. R>F 356-320 F>R 676 is subdivided into R>F
120-93 F>R 213 followed by  R>F 236-227 F>R  463, the latter being less
than 51% to over 49%. How would I value Falcon now in light of Muller's
findings, that Falcon + King mates as it does and the above statistics.
Initially Falcon > Rook until some number of pieces have been captured.
Certainly by 20 pieces on board Rook > Falcon. The valuations at different
stages of play by criterion of how many Pawns and pieces remain are in
flux, as more scores get studied. Falcon of course is never in general so much as 6.0
to Rook 5, nor so low as 4 to Rook 5. As my inquiry seeks clarification, I guess it is same style of viewing in this series so of course please leave it in.

Well, if we woud assume that the true score prcentage would be that of the
total run, i.e. 356/676 = 52.66%, the expected scores for the first 213
and last 463 games would be 112 and 244 points, repectively. So the actual
result is off by 8 points. This is kind of normal for 213 games: the standard
deviation over N games is ~0.5*sqrt(N), or 7.3 for 213 games. So the
observed deviation is indeed approximately 'standard'.

It is clear already from the 52.66 result that Rook and Falcon are very close matches, (although the difference from 50% is significant!), and that the details of the strategy (for which number of Pieces or Pawns would you prefer Falcons over Rooks, and for which the opposite) could improve the results much more than fine-tuning of a single Falcon value used during the entire game in the range R-0.5 to R+0.5.

Such subtle strategic evaluations are a bit beyond the scope of a simple program like Fairy-Max, which also doesn't use such considerations for the other pieces (e.g. it does not know about Rooks on open files, the Bishop pair or doubled Pawns). To do better, I would have to convert Joker80 to hndle the Falcon. (I already started with that, but I was a bit too busy recently to finish the job.)