Comments/Ratings for a Single Item

More like Warcraft.

the ferz knight in this interactive diagram is wrongly colored for black

the ferz knight in this interactive diagram is wrongly colored for black

Not for me. Try to flush your browser cache; this problem was already detected and fixed in Herculean Chess, and that fix should be effective here too.

This still is white for me.

It looks like a Black piece on my monitor.

Try adding "?nocache=true" to the end of the URL, press enter, then flush the cache with Ctrl-Reload.

ok, that does fix it

It is a bit difficult to diagnose the problem. By adding the ?nocache=true you bypass the CloudFlare cache, but the browser would also consider it a file that is distinct from the one without that suffix. So it would keep separate copies in the browser cache for each of those, and when you have seen the correct one through the ?nocache=true prefix requesting the file without prefix might still give you the obsolete one.

And the problem is that in the context of pages in articles by others you have no control over whether this suffix is added.

Whatever it is, it's not on my end. I've cleared the cache multiple times, tried different browsers and a different device even, and always see the white knightferz except with ?nocache=true (when viewing the image directly)

Hello!

I noticed a bug in the wyvern checkmate demo. For some time I even doubted that the Wyvern (ski rook) is a major piece, because there were some problems with the checkmating technique.

However, at the moment, these problems have been solved, the Wyvern (ski rook) is truly a major piece.

Here is a correct version of checkmate by the Wyvern (ski rook):

https://gifyu.com/image/SGMCu

The author, Dmitry Eskin, has updated this page.

Indeed, the Ski-Rook is a major piece. You can practice checkmating with it here.

If it only has a D move in one of the dimensions it is more problematic, whether it skips the first square in the other dimension or not. Because in that case it is bound to even or odd ranks or files. On an even-size board there than always is one edge it cannot attack, and if the bare King can reach there it has a fortress draw. If not, it is a forced win.

But the Ski-Rook can reach the entire board.

I am looking at the source code for the Interactive Diagram.

Can you explain why:

• Werewolf (Wazir + Mao) is WafsW?
• Unicorn (Ferz + Moa) is FafsF? I don't quite understand how 'afs' works. a = after, that's all I understood :)

How does the Interactive Diagram evaluate the value of pieces if it is not explicitly stated?

The a does not stand for 'after' but for 'and again', which is the same as 'before'. So afsW means W before fsW: first a W step in any of the 4 directions, and then another step at 45 degrees (fs) relative orientation to it.

Values are guestimated through a rather complex method, based on the empirical observation (i.e. from playing thousands of computer games starting from materially imbalanced positions) that short-range leapers with N moves have a value (33 + 0.7*N)*N centi-Pawn, where captures contribute twice as much as non-captures. To extend that to other classes of pieces their number of moves is calculated in a number of randomly generated positions with 25% filled boards, where the density of a certain color decreases to 0 at the furthest rank. And the average of this plus one standard deviation is then taken to plug into the formula for leapers.

Strength = Orthogonal directions. The primary role is played by the single vertical line directed towards the enemy - this line is much longer than the diagonals. The secondary role is played by the horizontal lines, allowing movement and attack of new verticals.

Speed = Diagonal directions. The key role is played by 2 frontal diagonal lines (right and left). Each diagonal line is shorter and weaker than the vertical one, but both can create several threats at the same time.

Dexterity = Knight directions. The key role is played by 4 frontal attack directions. Each such direction is expressed by a single attack point, and is much weaker than any line. But unlike lines, these attack points are numerous and unblockable.

(H) Hero = King

Humans = Speed + Strength:

• (s) Swordsman (versatile infantry) = Pawn, was Footman
• (C) Crossbowman (line shooter) = Bishop (B), was Monk
• (P) Pegasus (flying cavalry) = Knight (N), was Knight
• (E) Elephant (siege fortress) = Rook (R), was Griffin
• (T) Titan (heavy shooter) = Queen (RB), was Angel

Elves = Speed + Dexterity:

• (f) Fairy (light infantry) = Diagonal Pawn (fmceFifnmA), was Sprite
• (R) Ranger (long shooter) = Ski Bishop (jB)
• (U) Unicorn (light cavalry) = Fers + Moa (FafsF)
• (G) Griffin (siege flyer) = Leaping Rook-3 (WDH), was Pegasus
• (P) Phoenix (light flyer) = Archbishop (BN)

Orcs = Strength + Dexterity:

• (g) Guard (heavy infantry) = Orthogonal Pawn (fmceWsceW)
• (H) Harpy (flying shooter) = Fers + Alfil (FA)
• (W) Werewolf (heavy cavalry) = Wazir + Mao (WafsW)
• (C) Cyclop (siege shooter) = Ski Rook (jR), was Wyvern
• (D) Dragon (heavy flyer) = Chancellor (RN)

I'm thinking of returning all promotion options to the Guards, but slowing them down, taking away the possibility of an initial double move. This option needs to be tested in Fairy Max for balance. Orcs are a slow but powerful race, and slow Guards will only emphasize this racial trait.

With this update, the unit scheme will look like this:

Elves - Orcs:

https://www.chessvariants.com/index/listcomments.php?id=62339

There are several interesting things I have noticed while playing with different armies.

There are starting values ​​of pieces that were previously selected by replacing them with other pieces and comparing the chances of winning in a series of Fairy Max games. The standard values ​​for comparison were 3.25 / 3.25 / 5 / 9.75 evaluations (knight / bishop / rook / queen). In fact, these are exchange values ​​that are far from the starting values, and the standards themselves should be revised (calibration procedure).

And there are exchange values ​​of pieces during the game (in the opening, middlegame, endgame). And all these values ​​can differ.

I found that the archbishop is much more active and dangerous in the opening than the chancellor or the queen. Obviously, it becomes much weaker towards the endgame. But its starting value is quite high.

What's even more interesting is that the Unicorn or Werewolf is weaker than the Knight in many ways, especially in the opening. Despite having more attack squares (12 vs. 8), they don't jump, and can very often be blocked and trapped. Regular Knights can easily jump ahead of the infantry, and retreat if necessary. But the Unicorn and Werewolf have a lot of trouble both getting into the game and retreating through the pawn fence. And it's especially cool that each piece has its own strengths and weaknesses, and not that they are "extended" knights with only advantages.

I also noticed that the line rook "in ideal conditions" is indeed much superior to both the jumping rooks of the Orcs and Elves. But these "ideal conditions" do not appear for the line rook soon, and the jumping rooks pose a significant threat already in the opening. All this means that the exchange value of a line rook can be much higher than its starting value, which is approximately equal to both jumping rooks of the orcs and elves.

What is also interesting is that the elven bishops can make the strongest diagonal attacks, since any exchange for a piece is advantageous to them, and it is not always possible to close with a pawn due to a jump. Orc bishops are interesting for their unusual forks, which are impossible to close (as well as from knights). The value of both of these bishops is high in the opening, but drops sharply closer to the endgame.

I would like to not only get the exact starting values ​​of all these pieces, but also use these values ​​to calculate their exchange value in the opening, middlegame and endgame.

Any help in this direction is welcome. :)

Before the use of Neural Networks for evaluation became common, the strongest chess programs all used two sets of piece values (as well as for some other parameters, like passed-pawn bonuses, mobilities...), referred to as 'opening values' and 'end-game values'. Two evaluations are then made based on these, and the actual evaluation is then calculated by linear interpolation of these as a function of 'game phase'. The latter is based on the (weighted) number of pieces present, e.g. 4*queens + 2*rooks + minors or 6*queens + 3*rooks + minors. (Pawns are usually ignored; if anything these should probably have negative weight.)

Systematic increase of piece values towards the end-game, without changing their ratio, would encourage even equal piece trading when ahead (to advance the game phase).

Fairy-Max doesn't have such an advanced evaluation. But in my experience moderate misconceptions about the piece values are not affecting the outcome of material-imbalance games much. As long as both players share the same misconception.

To get piece values for later in the game, you could start material-imbalance games from early end-game positions. E.g. with each player 4-6 pawns in a symmetric setup and 1-3 (non-royal) pieces. This is a bit more tricky, since the shorter remaining game duration leaves less rooms for the players to make errors. So an extra Pawn might already be enough to produce a 100% win rate, making it impossible to estimate how large the advantage actually was. randomizing the first few moves might no longer make it possible to get a little more spread-out results. I guess this could be cured by averaging over a large number of different, not necessary symmetric pawn structures used as starting positions.

I once did a test for Archbishop vs Rook + Knight as only pieces in the presence of 4-6 Pawns, and the Archbishop still came out on top. Even with an extra Pawn for the R+N it could often beat them. So I am skeptical about the Archbishop value going down in the end-game.

The anomalously high value of the Archbishop seems for a large part a result of its power to destroy Pawn chains. This might be coupled very much with how Pawns move. In general piece values are also dependent on the detailed material composition of the opposing army, rather than just on a global measure of its total strength. E.g. 3 Queens vs 3 Queens (in the presence of equal Pawns) would obviously be equal, but 3 Queens lose very badly to 7 Knights. Despite the fact that conventional piece values predict 3 Queens to balance 9 Knights.

So that you observe the value of Archbishops to go down towards the end-game might also be due to using it in a different-armies setting, with different Pawn types. Perhaps it loses the effectivity that is exhibits against FIDE Pawns, and this would hurt more in the end-game. Where the ratio Pawns vs pieces usually goes up. (The stronger the pieces, the earlier they usually get traded out of the game.)

It's not like the Archbishop's value drops off that much later in the game. He's similar to the Knight in that regard.

But the way it seems to me:

In the opening:
Knight > MoaFerz or MaoWazir
SkiBishop > AlfilFerz > Bishop
Rook < SkiRook or FlyingRook
Knight > Bishop > Rook
Archbishop > Chancellor > Queen

In the middlegame:
Knight < MoaFerz or MaoWazir
SkiBishop < AlfilFerz < Bishop
Rook = SkiRook or FlyingRook
Knight = Bishop < Rook
Archbishop < Chancellor = Queen

In the endgame:
Knight < MoaFerz or MaoWazir
SkiBishop < AlfilFerz < Bishop
Rook > SkiRook or FlyingRook
Knight < Bishop < Rook
Archbishop < Chancellor < Queen

The starting value of the pieces is determined by the sum of their influence in the opening, then the middlegame, then the endgame (so knight = bishop).

But the exchange value of the pieces is determined at the moment of their exchange, that is, usually it is the sum of their influence in the middlegame + endgame (so bishop is slightly stronger than knight).

I would like to check all these estimates with the help of precise digital estimates of Fairy Max, but it will take a lot of time anyway. The maximum speed that can be achieved is 6 thousand games per day (24 hours), which gives an error of about 1% in the win rate (or less than 10% of the pawn).

Another thing I'd like to find out:

Does SkiBishop really have such a low starting value as a 2/3 Bishop? Or is it because he was in a worse position on b1/g1 when he was tested (it was many years ago, I don't remember exactly)?

Perhaps MoaFerz is slightly stronger than MaoWazir precisely because of his position on c1/f1 (he currently has a classic position on b1/g1)?

Although MoaFerz has a lot of advantages:
- MoaFerz has checkmate potential, while MaoWazir does not
- MaoWazir must alternate square colors, while MoaFerz does not
- MoaFerz is generally more active than MaoWazir, since Ferz is faster than Wazir, and has 2 frontal directions instead of 1

However, MaoWazir has the following advantages:
- MaoWazir is stronger than MoaFerz on the edge of the board, since Wazir has more moves there than Ferz
- MaoWazir attacks more new squares on its move than MoaFerz (alternating square colors gives this advantage)

How does the balance of power change between SkiRook and FlyingRook? FlyingRook has a double jump, so in theory it should have an advantage in the opening. The SkiRook often behaves like a line piece, so it should ideally have an advantage in the endgame. However, both of these Rooks should eventually tend towards 2/3 of the Rook, by which time their jumps will no longer matter due to the small number of pieces.

Another interesting thing is that Interactive Diagram often trades SkiRook for MaoWazir in the second half of the game. Perhaps the trade value of these pieces is really close, like Knight and Bishop? Despite the fact that SkiRook's starting value is a whole pawn higher.

Personally, I noted that closer to the endgame I tend to trade the Skirook for the FlyingRook. Despite the fact that the SkiRook has a linear component, it feels like this figure is more static and less convenient to move than the FlyingRook (while in the endgame mobility is especially in demand, since individual pieces move more often than in the opening or middlegame). It has a weakness in close combat, which is easier to exploit than the FlyingRook's lack of range. Even checkmating with FlyingRook does not cause any problems, but in case of SkiRook you need to learn (the difficulty is close to checkmating with Bishop and Knight). If the Rook or FlyingRook moves to take a stronger position, many of the SkiRook's endgame moves are simply trying to hold the line by eliminating a near-field weakness.

Bishop > Rook might be true in the opening if you look at the instantaneous tactical power. But most pieces will survive into the late middle-game, where the value will reverse. So the Rook, even when pretty useless, has a lot of 'latent value', which will be realized automatically as the game phase progresses. This is especially true if both players have approximately the same latent value; then the difference in instantaneous tactical value is not so large that the game can be decided before the latent values are realized. For this reason piece values are usually dominated by the end-game values, and starting with a Knight instead of one of the Rooks is an almost certain loss. It might be different if one player had 7 Rooks and the other 7 Knights, although even then I doubt the Knights have so much instantaneous advantage that they could quickly decide the game. The Rook might not have much offensive value early in the game, but it can still be useful for defense. 7 Rooks + 1 Pawn would probably lose against 7 Knights + 8 Pawns, though.

In games with piece drops, like Shogi or Crazyhouse, you basically are in a permanent middle-game situation, and the instantaneous middle-game values become relevant. Indeed Shogi players value Rook and Bishop about the same.

I meant that the bishop is stronger than the rook specifically in the opening stage, and not in terms of the cumulative starting value (which includes the opening + middlegame + endgame). If you need to gain an advantage specifically in the opening, the bishop will do more in this regard than the rook.

In other words:

Starting value = Opening value + Middlegame value + Endgame value (since opening value is present, the starting position plays a role).

Average exchange value = Middlegame value + Endgame value (since pieces are usually exchanged in the middlegame).

I largely agree with that. Except that the averages probably have to be weighted, with a significantly larger weight for the end-game. This assumes the game will be decided in the end-game, though.

This issue of latent values is a really complex one, as in principle the weighting is affected by the probability that you will actually reach that game phase. If the difference in middle-game values is so large that the game can be decided before you reach the end-game, the end-game values would obviously carry no weight. In Chess the latent value of pieces is not so large that having one or two pieces that perform better in the middle-game compared to their end-game value would be a decisive advantage, though.

That is different in Chu Shogi, which is Chess-like (i.e. no piece drops), but where all pieces can promote, in a 4-rank-deep zone. So for most pieces it will eventually become impossible to prevent their promotion, and the end-game value approaches that of the promoted form. For some pieces this can make a huge difference with their middle-game value, where they would not be promoted: some 8-target leapers, which should be of Knight value, can promote to pieces stronger than a Queen. If you judge those by their end-game value you can get in a situation where the opponent has an army with a middle-game value that is so much larger than yours, that he can start promoting his pieces long before you can, driving up the instantaneous value difference even more, and use his ever-growing advantage to make sure you never get to promote anything.

Off topic a bit. I wonder if the spartan chess spartan army can be added to this game (Dark Elves?). I think we can safely assume that it is as strong as the human army. But what about the other two?

I'm potentially looking into adding new armies, but would like some figure ratings for the ones I have.

I will add the exchange values ​​of the pieces (without confirmation from Fairy Max yet):

Humans

• 1.0 Swordsman = Pawn
• 3.75 Crossbowman = Bishop
• 3.5 Pegasus = Knight
• 5.25 Elephant = Rook
• 10 Titan = Queen

Orcs

• 1.25 Guard = Slow Orthogonal Pawn
• 3.0 Harpy = Flying Bishop / FerzAlfil
• 4.5 Werewolf = Orthogonal Lame Knight / MaoWazir
• 4.5 Cyclop = Ski Rook
• 9.0 Dragon = Chancellor

Elves

• 1.5 Fairy = Diagonal Pawn
• 2.5 Archer = Ski Bishop
• 4.5 Unicorn = Diagonal Lame Knight / MoaFerz
• 4.5 Griffin = Flying Rook / WDH
• 8.0 Phoenix = Archbishop

These evaluations differ from the opening evaluations, since the exchanges usually take place in the middlegame.

Here is an Interactive Diagram of alternative icons for Elves and Orcs (example).