[ Help | Earliest Comments | Latest Comments ]

[ List All Subjects of Discussion | Create New Subject of Discussion ]

[ List Earliest Comments Only For Pages | Games | Rated Pages | Rated Games | Subjects of Discussion ]

Comments/Ratings for a Single Item

⇧Earliest ⇧Earlier ⇧Reverse Order⇩ Later⇧ Latest⇩

0000000100000000[Subject Thread] [Add Response]

Aurelian Florea wrote on Wed, Sep 14, 2016 04:30 AM UTC:

I personally don't like multiplayer variants as I consider most of them unfair from the point of view that multiple players may attack the same victim. So if that's up to me I won't consider those. Playing against humans of the same level or close it's a good idea but it assumes a large enough number of players, not the case at least initially.

Regarding authority for the game, well the programmers are the main authority, who else. We or they (I'm not sure yet if I will get involved as I am more for more challenging tasks from an intellectual point of view) would consult the community of course.

Time should be enough maybe at blitz or a bit longer, not too fast as the player is not used to playing most of the variants, but not too tedious either.

Aurelian Florea wrote on Wed, Sep 14, 2016 05:05 AM UTC:

I was concerned about the opening in the second game as I put the queen and marshall far from the center of the board. That happened mostly because I was not use to with an archbishop in the middle, but after a few trials it seems fine. I have seen the archbishop in the middle in Paulovich's variant of Capablanca chess, so I guess I am not the only one with this problem. I have done this in order to have the minor pieces (the ones who acording to opening principles attack first) closer to the middle. Although bishops could be 1 square more lateral.

H. G. Muller wrote on Wed, Sep 14, 2016 07:16 AM UTC:

The submission software now also general posting of HTML in comments. That means it is also possible to put an interactive diagram there. But as you can see from this example, the formatting is not really adapted to that. (*)

files=3 ranks=4 holdingsType=-1 promoOffset=3 graphicsDir=http://www.chessvariants.com/graphics.dir/alfaerie/ whitePrefix=w blackPrefix=b graphicsType=gif squareSize=54 symmetry=none chick::fW:pawn:b2,,b3 elephant::F:elephant:a1,,c4 giraffe::W:giraffe:c1,,a4 hen::WfF:dragon:, lion::K:lion:b1,,b4

The diagram script is not really suitable for that either, as it is not possible to have more than one interactive diagram per page. The JavaScript searches for a HTML element with id="diagram", takes the diagram description from there, and puts the HTML code to display the diagram in its place. If there would be multiple elements with this 'id', it would only see the first. This is not easy to fix. And it means that people can pre-empt each others diagrams. You can always click 'View', however, to see a posting in isolation.

So perhaps this practice should not be encouraged.

[Edit] *) When I removed the floating style of the diagram it is not so bad, and the diagram no longer extends into the preceding posting.

Aurelian Florea wrote on Wed, Sep 14, 2016 10:46 AM UTC:

I tinkered those 2 games a little bit and I came up with the following value of pieces based on their values in Omega Chess (omega chess website), Grand Chess (H.G.Muller values from wikipedia), Ralph Betza's evaluation of the aanca and griffin, mobility calculation for zebra, and my own observations playing the games, I also named the two games so I think I'm ready to maybe submit them, eventually:

Apothecary chess1

rook 6

wizard 3.8

champion 4

knight 4

bishop 4

queen 12

aanca 7

griffin 8.75

Apothecary chess2

zebra 3

camel 3.2

elephant 3.7

knight 3.4

bishop 4

queen 12

archbishop 10

marshall 11

I decided to name the games apothecary chess as they are obtained adding somewhat arbitrary mutations to well known games in the way apothecaries used to tinker their receipts.

H. G. Muller wrote on Wed, Sep 14, 2016 11:37 AM UTC:

I am surprised by the values you assign to Aanca / Gryphon. They seem too low. I measured only on 8x8, and IIRC on the Kaufman scale (minor = 3.25, Q = 9.50) the values I got were Aanca=7.8 and Gryphon=8.3. And as these are sliding pieces, they should benefit just as much from larger boards as Rook and Bishop. An Aanca covers more squares than two Bishops (positioned left and right of it) would, so I would expect its value to be at least double that of the Bishop. It is also not color bound.

Because of the up-shifted starting position the board of tehse Grand-Chess-like games is effectively 10x8, and wide boards benefit diagonal movers more than orthogoal movers. (This because the chance that both their forward moves hit the enemy increases.) I would not be extremely surprised as an Aanca on a cylinder board would be worth more than a Gryphon.

Aurelian Florea wrote on Wed, Sep 14, 2016 12:09 PM UTC:

Actually my mobility program gives the griffin roughly equal to a queen a little bit weaker on crowded boards but even stronger in the endgame. The Aanca is the strongest among the 3 in the beginning but quickly loses power becoming 21.5/26 queens in the endgame. Clearly stronger than stated. I think the conondrum comes from the fact that mister betza calculations are based on rook and bishop. So the result is that the rook and the bishop are stronger. The bishop is still 4 (ok maybe 4 and something) because it gets a colourboundness penalty, but the rook must be higher. My mobility program give 0.63 queens for the rook.

That beeing said 4 pawns are definetly better than a bishop in the endgame. Remember the rule that they can even promote to a rook on rank 8, but I never got that far with my calculations.

So new values:

rook 7.25

wizard 3.8

champion 4

knight 4

bishop 4.2

queen 12

aanca 9

griffin 11

Apothecary chess2

zebra 3

camel 3.2

elephant 3.7

knight 3.4

bishop 4.2

queen 12

archbishop 10

marshall 11

rook 7.25

Aurelian Florea wrote on Wed, Sep 14, 2016 12:12 PM UTC:

The html code posted earlier doesn't enforce the rule that the pawn can promote to almost a rook on ranks 8 and 9 so you must decline manually not to promote to an aanca, griffin, queen , archbishop, marshall.

Aurelian Florea wrote on Wed, Sep 14, 2016 12:14 PM UTC:

the archbishop is a bit high too maybe!

Aurelian Florea wrote on Wed, Sep 14, 2016 12:24 PM UTC:

Actually I don't see many variants with aancas, griffins, or zebras. I assume there is good reason for that. That's why I considered the name paranoid chess, too, but my father advised me against it!

H. G. Muller wrote on Wed, Sep 14, 2016 12:41 PM UTC:

> The html code posted earlier doesn't enforce the rule that the pawn can promote to almost a rook on ranks 8 and 9 so you must decline manually not to promote to an aanca, griffin, queen , archbishop, marshall.

It also does not allow you to refrain from promotion: there is no 'P' in the promoChoice.

If you really want this to be handled automatically, you can add the following text to the HTML:

<script>function WeirdPromotion(x1, y1,x2, y2, promo)
{
  if((board[y1][x1] & 15) != 1) return promo; // moved piece is not a Pawn
  if(y2 == 9 || y2 == 0) return ((promo & 15) == 1 ? promo+4 : promo); // on last rank: cannot stay Pawn
  if((promo & 15) > 4) return board[y1][x1]; // larger than Rook: remains Pawn
  return promo; // choice was acceptable
}
</script>

In Capablanca Chess A + P is slightly stronger than Q, and I don't expect that tobe much different here.

[Edit]The promotion code above assumed a piece ordering P N B R Q ..., so that 1 is Pawn, 4 = Rook and 5 = Queen. I see now that you did not define the pieces in that order, and you would have to adapt that,or the code.

Aurelian Florea wrote on Wed, Sep 14, 2016 12:46 PM UTC:

I appologise to have forgotten about this function, I've done so many attempts yesterday evening with other matters.

Aurelian Florea wrote on Wed, Sep 14, 2016 01:03 PM UTC:

I had to change the order in promochoice, but the function works fine provided the piece list is open, or so I think, but that's not an issue

Aurelian Florea wrote on Wed, Sep 14, 2016 01:08 PM UTC:

So no these are the scripts that allow you to play the games until in a few days when I post them on CVP!

H. G. Muller wrote on Wed, Sep 14, 2016 01:17 PM UTC:

Indeed, that is a disadvantage of the way promotions are implemented: you have to select from the piece list. When used here as diagram the intended application was to show pseudo-legal piece moves, not play complete games, so promotion is really a bit outside the scope. When I use it for playing games on a turn-based server page, I display the piece list permanently next to the board, so that this way of chosing the promotion piece is not a problem. (And for Shogi variants, where the list can get very long, the choice is only yes/no, and can be made on the line above the board.) I guess it could also open a list of names to click on.

I noticed the diagram allows you to promote to opponent pieces! For some reason it does not allow promotion when moving from 8th to 9th rank after deferral. I should certainly fix that.

I guess you can do without P in the promoChoice, so that clicks on the Pawn will be ignored, as you can defer from promotion by clicking a piece that is not allowed on the 8th rank. That would avoid assignment of a default promotion piece when you choose Pawn on 10th rank.

Note that you have an M in the promoChoice string, presumably for Marshall, but that the piece name is defined as Chancellor, abbreviated by C. So the choice of Chancellor would be ignored.

PS: note I made a tiny change to the WeirdPromotion script, as when you chose a Pawn for black on the last rank you would get a WHITE Queen!

Aurelian Florea wrote on Wed, Sep 14, 2016 01:19 PM UTC:

Thanks, I'll fix that!

Aurelian Florea wrote on Wed, Sep 14, 2016 01:50 PM UTC:

It doesn't work properly in the 3rd rank it promoted only to QBN it is suposed to be other pieces, and yes it does not promote at the 9 th rank. Don't worry though. It works fine when hadled manually!

H. G. Muller wrote on Wed, Sep 14, 2016 02:24 PM UTC:

The WeidPromotion script says that promotion to pieces > 4 should be refused (when not on last rank). I wrote itthat way, because I assumed the order P, N, B, R, Q, ...But you defined the piece in the order P, B, R, Q, ... and the Knight last. So then only B, R and Q are <= 4. You should edit the definition of the diagram toorder the piece definitionlines such that allpieces you want to allow promotion to on 8th/9th rank immediately follow Pawn, and put the forbidden pieces behind them. And then you might have to adapt the 4 in the WeirdPromotion script to the size of the allowed subset.

Kevin Pacey wrote on Wed, Sep 14, 2016 02:28 PM UTC:

Aurelian wrote:

I was hoping to yet hear from Fergus on what he had in mind on this particular matter. If he foresees a corporation one day making the effort, it wouldn't be the programmers, as they would just be instructed what particular games to program. If Fergus had faster progress in mind, it might be a programming effort on the part of some CVP community members, or perhaps some outside programmers with free time who are hobbists (might they be invited &/or instructed by CVP editors, in such a case?).

Aurelian Florea wrote on Wed, Sep 14, 2016 02:41 PM UTC:

Ok, HG. I understand!

Aurelian Florea wrote on Wed, Sep 14, 2016 02:59 PM UTC:

I did it correctly once again this is how it goes:

🕸Fergus Duniho wrote on Wed, Sep 14, 2016 03:10 PM UTC:

I was hoping to yet hear from Fergus on what he had in mind on this particular matter. If he foresees a corporation one day making the effort, it wouldn't be the programmers, as they would just be instructed what particular games to program. If Fergus had faster progress in mind, it might be a programming effort on the part of some CVP community members, or perhaps some outside programmers with free time who are hobbists (might they be invited &/or instructed by CVP editors, in such a case?).

I proposed it as a way to structure a tournament. I think it was H. G. who suggested making software along these lines. If software were made, it would be in the hands of those who are in charge of the software's design, whoever they are, and I wouldn't have any say in this unless I was the person in charge.

Kevin Pacey wrote on Wed, Sep 14, 2016 03:25 PM UTC:

Sorry Fergus & H.G., I thought I'd at least remembered it was Fergus who suggested a video-game style way of increasing interest in chess variants. In such a case, I should note it might be a good idea to have slick graphics & soundtrack before and after each game is played, and one member of a programming team might be dedicated just to this part.

Aurelian Florea wrote on Wed, Sep 14, 2016 03:35 PM UTC:

I think both the tournaments, and the video game ideas are good, should we start a post on each of those?

Aurelian Florea wrote on Wed, Sep 14, 2016 05:17 PM UTC:

H.G.,

About the pawn, this pawn is slightly weaker than in capablanca as it takes slightly longer to promote, not a big deal, if it's a passed pawn will eventually queen. But there is also a strengthening effect as it can promote at rank 8 to rook (well to a plethora of pieces but the best choice is usually rook). I have no clue how to evaluate this effect but it shouldn't be much as early under promotions will not happen often, I raise that matter because it's specific to those two games.

On the matter of archbishop+pawn there is no reason to doubt you, but I honestly don't know how to normalize everything. There is also no reason to think queens have different strengths between to two games for example. and what about the situation (admittedly rare) R+3P VS archbishop, point being is hard to find a balance.

I'd be honored if you take your time and put forward some values of your own, or on the other hand put some values at first glance.

I think in these cases machine learning could tune very well values, I've read an article somewhere on atomic chess and other games but I can't find it now.

And it doesn't have to be a 1dimesional value an archbishop could worth 10.5 in an A+P vs Q ending and 9.5 in a R+3P vs A ending. I don't know.

You were correct on yesterday assessment that the game is tedious. It takes a lot of time to reach endgame, actually in all cases I ended games through small blunders of pieces in early or late middle game. It became obvious IMO who'll win. :) At least it's not Taikyoku shogi I doubt anyone started that one, finishing it is out of question I guess.

H. G. Muller wrote on Wed, Sep 14, 2016 06:25 PM UTC:

I once measured the effect of restricting promotion to a minor for one side in Spartan Chess, to see if this was a good way to tune the balance. (After all, players have 8 Pawns there, so even if it only very slightly affects the value of a Pawn, it counts for a lot in the strength of the army.) But I could not measure any effect of it. I guess that in practice promotions can almost always be prevented by sacrificing a minor for the Pawn, and that where they can not, they would be decisive whatever you promote to.

Based on that observation it should not even matter much for the Pawn value if promotion was restricted to Rook everywhere. In which case you would of course always promote to Rook on the 8th, as nothing is to be gained by postponing it. So promoting to Rook must be a viable strategy, and the opponent must defend against that. So most of the time he would sacrifice a minor for the passer even before it reaches 8th rank, or he would make sure he can capture it when it reaches 8th rank, not to give you a free Rook. The only reason not to promote to Rook would then be that you can see you can force it to the 10th, in which case the opponent would also sac a minor for it when it reaches the 8th as a Pawn.

In the end-game, in the case of a promotion race, you cannot afford to delay promotion until you reach the 10th, as the opponent would then promote to Rook on his 8th, and a Rook would be enough to stop your passer completely. So promotion on the 10th becomes mostly theoretical, only occurring in cases that were one side is already completely lost. So the only factor effecting Pawn value is that promotion is shifted up to the opponent Pawn rank, rather than a rank behind it as in orthodox Chess. This should increase the Pawn value. Also because KPK now always is a win.

25 comments displayed

⇧Earliest ⇧Earlier ⇧Reverse Order⇩ Later⇧ Latest⇩

Permalink to the exact comments currently displayed.