Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - joejoyce

Pages: 1 2 3 »
« on: October 07, 2018, 11:23:30 am »
I come for interesting conversations and ideas, to talk design more than tactics and strategy for one particular game - unless it is representative of a particular type of variant. Like many, I will push my own ideas, but unlike some, I won't only be here to push my game. Make it interesting for more people like me and HG to participate.

"With the lights out, it's less dangerous
Here we are now, entertain us
I feel stupid and contagious
Here we are now, entertain us"

Variant Theory / Re: Maths question: How many possible setups...?
« on: September 29, 2018, 09:58:55 pm »
Let's start with rooks. How many rook combinations are there? There are 2 rooks, each of which can be 18 similar but not identical pieces. SO there are 18x18=324 possible combinations for rooks. This number will be the same for the knight and bishop pairs, so that will give you 324x324x324 possible combos for the B, N, R pieces. Since the king is always just a standard king, it's multiplicative value is 1, and the queen's is 18. So the back rank is 324x324x324x1x18.

Okay, now the pawns; each pair of pawns gives us 324 possibilities, and there are 4 pairs, giving us 324x324x324x324, which will multiply with the back rank to give the answer. Since so far I'm doing this in my head, I'm keeping the numbers simple... 324x324=90000+12000+2400 (300x300+2x20x300+2x4x300) = 104.400, using European notation, or 104,400 for US notation.

Now I'm gonna cheat! call 104400 ~= 100000, so the pawns have ~100000x100000=10000000000 possibilities, roughly (it's a little low). Now, the back rank is 324x324x324x18, or 100000x324x18 ~= 100000x300x20=600000000, giving a final estimate of  10000000000x600000000=6000000000000000000, or 6x10^18, and this is a little low.

Grin, somebody want to check my numbers? It's been a long time since I did probabilities!

General Discussion / Leela vs. Stockfish9
« on: September 17, 2018, 10:49:03 am »

General Discussion / Re: Forum Announcement and Contest
« on: June 08, 2018, 05:42:17 pm »


   King: K - standard chess/shatranj king, moving 1 square in any available direction.

   Knight: N - standard chess/shatranj knight, leaping 1 square orthogonally and then one more square diagonally outward.

   Modern Elephant: E - Alfil + Ferz; steps 1 square diagonally or leaps 2 squares diagonally.

   Modern Warmachine: W - Dabbaba + Wazir; steps 1 square orthogonally or leaps 2 squares orthogonally.

   Shatranj Pawn: P - standard chess pawn except with no initial double-step. Promotes to Knight, Modern Elephant or Modern Warmachine.

Game is played on a 5 x 6 square board. At start, each player has 5 pawns, placed on the 2nd and 5th ranks, and 7 pieces, placed along the 1st and 6th ranks.
    Piece placement: w+n; e; k; e+n; w / ppppp / 5 / 5 / PPPPP / W+N; E; K; E+N; W

Ralph Betza's Rule Zero applies: All standard chess rules apply except as otherwise noted.

Any 2 friendly pieces may occupy the same square. This does not apply to pawns (unless both players agree to try this experimental version.)
This pair of pieces may act as 1 piece, moving, capturing and being captured together, as either of the pieces in the pair for each move.
Or either piece may freely move as an individual, leaving the other piece behind on the originating square.
A piece may start its move as 1/2 of one pair, and end that move as 1/2 of another pair.

I've played a few games of shatranj, but far more shatranj variants. Have you considered looking at all the more modern variants of the game that retain short range and leaping pieces?

If we can use already designed games, here's The Battle of Macysburg, a Command and Maneuver scenario. It's an abstract riff on the American Civil War battle of Gettysburg, a multi-move shatranj variant featuring 'mini-kings' which lead small groups of pieces across the board. The game lasts 38 turns, divided into 3 'days' of (about) a dozen turns each, separated by 2 'night' turns where the armies break off combat and withdraw for the night to rest and rally troops. Here's the first day and night, with new units coming on to the board marked: While I don't believe it's noncomputable, I do believe it's very difficult and I suspect a small group of human experts playing against Alpha Zero might well hold their own. Macysburg has been playtested enough to show it works fairly well. There are some minor modifications to the  turn structure I'd like to try out, specifically changing the length of a day from 12 turns for each of the 3 days to 13 on day 1, 12 on day 12, and 11 on day 3, to represent army fatigue and to allow the first day's reinforcements (2nd wave of units) to come in 1 turn later, to (try to) encourage the initial units to come into contact with the enemy without waiting for the day 1 reinforcements to catch up before joining battle. The minor variability of the turn structure is just one of the variable features of this game. The terrain is variable, the entry points and times are variable, the make-up of each of the 4 waves of units entering during the first 2 days of the battle, all are variable. And the 'movie' gives you an idea, but nowhere near a full picture of how predictable the future might be in the game. It's the 2nd best game I can offer for this contest.

Correspondence Game Requests / Re: Game Courier Gamerooms
« on: May 07, 2018, 12:27:49 pm »
Here are some games that I'm inviting this forum to.
Did you get opponents for your games? As far as I know, you're posting in the only place where you have a real chance of finding opponents.

Variant Theory / Re: Experimental large multi-move variant.
« on: May 03, 2018, 12:19:23 pm »
I made a preset for this game on Game Courier. Are you a member at If so, send me an invite. If not, join, then send me an invite: joejoyce

As far as improving chess, I don't think any one game will do it. The only real improvement I see is an ever-expanding list of high-quality chess variants and fusions of chess and other games. The major regional chess variants on Earth now are all rather polished games that show strong signs of a common origin followed by different development. Xiang Qi, FIDE, and shogi are all excellent and worthy and different games. And that is the way 'forward', multiple games. Like sports has individual sport championships but also triathlons, decathlons... chess needs an expansion of the games it plays to grow, in my opinion. I'd like to see chess championships in 10 games randomly picked from the 1000 'best' chess variants and 5 random games from the top 100 as well as the champion of the 'Top Three', Xiang Qi, FIDE, and shogi. So I'd say the best thing you could do would be to find, design, play, and publicize the best variants you can.

...Did you look at the variant I just posted?
Around 2009, I played Simplified chess face-to-face with Rich Hutnik at Spielbany. Forgot whether the pawns moved 1 or 2 on first move. I remember it as a very different game for experienced chess players, kinda dirty and gritty. I find it's an excellent teaching game, if you wish to understand a bit more about how things in chess work, and why. I also recommend pairing it with the 'opposite' variation, playing on an 8x9, with 5 empty rows between pawns, and all chess rules as standard, except of course for the square-color placement of the black pieces, which becomes the same as the white pieces.

I'm one of the few who think varying the board can easily be as important as varying the pieces. And I'm also a big fan of shatranj pawns, with no double step. I've noticed few moderns make many pawn moves in modern shatranj variants. While I may make too many, most don't make enough, because they are too used to the diminished role of pawns in the modern game, and don't appreciate they are actual fighting pieces in shatranj. But I do see why the pawns get an initial double step in the modern game. The modern pieces are essentially overpowered, with 5 of 7 non-royal pieces infinite sliders and the other two leapers. Without that double step, the pawns are reduced to speed bumps. Of course, reducing the number of empty ranks mitigates that effect.

Variant Theory / Re: Experimental large multi-move variant.
« on: April 30, 2018, 06:50:24 pm »
Interesting try. I don't recall seeing this design, a direct expansion of each piece and board square into 9*, although I've run across somewhat similar ideas. Is it playable online? I'd help playtest it.

As far as how it will play, let me first steal a line from my recent comment to HG in another thread. "... the 'piece' now has "hit points", and can survive losing some units. It has general attack and defense abilities that are based on the number of units in the piece and how the units move, modified by the *exact*placement of all units involved..." *except the king. You use a 'crippled king' strategy here. Any reason why, except you didn't want to play 'capture the king'?

You use rigid 'pieces' that can partly interpenetrate to make up for their rigidity. But, as your example game shows, you lose parts of 'pieces'. That changes the strategy of the game somewhat from standard chess, because as the game wears on, players tend to use the pieces with the most remaining units - this observation based on experience with chesimals, which are essentially a flexible version of your 3x3 square piece.

Further differences from standard western chess include GESS-like creation of pawns, and 'color-changing' bishops, or at least bishop units changing from one 'army' to the other, as apparently any 3x3 set of board squares can define an army. That leads to an even greater fragmentation of 'pieces'/armies into smaller clumps of similar units. The armies become disorganized. In fact, attacking to disorganize your opponent's armies becomes a viable strategy if you can develop one or more ways to do that.

I suppose you could restrict the 3x3 squares to the 64 defined by the initial set-up of 'back rank pieces' but this would remove some/much of the fun and strategy. How about you have a "rally" command that calls back scattered units of one type to re-form near the king? Maybe call back all units of 1 type in a player-designated 7x7 square, for example?

Anyway, is there a way to play this online?

I don't see how multiple moves per turn shortens the game. It will decrease the number of turns, for sure, but you still have to make all the moves. It only changes the order of the moves.

If you have very many pieces, it seems a big problem to have them all participate in a meaningful way, and still keep a reasonable game duration.
I think to a great extent, time taken is dependent upon the coordination of the moves. If each individual move is totally independent of what is going on elsewhere, then you are correct, and there is no difference in time to play. However, if there is coordination among piece moves, then the calculation time, in general, tends to drop. Yes, you may envision some bizarre, far-reaching group of complex moves that decimates your opponent and gives you far and away the better position, and this might well take more time, but how often can it happen realistically? And if that is the case, the future of the game being played will be shortened by that move, in general. I grant exceptions, but they are just that, exceptional, and not something you would deal with often.

If you use activators, then the time almost has to be shortened, because you have a tight-knit group of pieces all moving together in the game, even though they move one at a time on the board. Basically, all the other pieces have to go along with the activator, or they won't be able to move next turn and thereafter, until they are picked up by an activator again.

What are the benefits of a multiple unit activator "piece", besides ease of movement? Well, the 'piece' now has "hit points", and can survive losing some units. It has general attack and defense abilities that are based on the number of units in the piece and how the units move, modified by the *exact*placement of all units involved in or near the action. Now consider the last example I gave, a 50x50 with 40 3x3 9-unit activator pieces. I would argue that just on the face of it, the 40 *piece* game is much more likely to move more of its *units* than the 360 piece game, especially if the game is played one move per player turn.

I grant that it's obviously true that a single move/turn game where most pieces are ignored, and all the action is concentrated in a few pieces near the kings, can end faster than a multimove version of the same game, but there are 2 points here. One is yours - why bother with a large game when you are only actually playing a small game? Exactly! There is no reason, other than to add confusion (which may be a valid reason on rare occasions.) The second point is mine - the multimove game provides a far fuller and richer game experience as well as using far more of the individual units on the board. That is, if the game is properly constrained by the rules. You could set things up so that white always wins on the first move, but that isn't a game I'd play very often.

Finally, very large chess variants can exhibit more different behaviors that the standard games. Macysburg is a riff on the American Civil War battle of Gettysburg in 1863. Heh, and I was actually playing with a chesimals game that evolved chesimals in an 'ocean' (puddle) over the course of time. I'd gotten up to multiple organ chesimals with offensive and defensive capabilities, and also the need to 'eat', to harvest 'plants' in the ocean for raw material to build, repair, and mutate during a game. Then Nick Bentley announced a new game from the company he now works for that covers similar territory in what sounds like a much simpler way, so I shelved that project for the near future, at least.

I've tried to push the boundaries of what chess can do with a lot of my games. So in many ways, I'm conservative, in an attempt to tempt people to take just one (sometimes large) step from chess to whatever strange thing I dreamed up, instead of a bunch of steps. I design games I'd like to play, so they have to be able to attract the occasional player. Creating your own universe where everything is different and novel creates an almost insurmountable learning curve for new players, giving very few opportunities to actually play. The pile of short range pieces I've designed is introduced in 'shatrangized' versions of Modern chess, Capablanca's/Carrera's Chess, and Grand Chess. Grin, that way, they actually get some play! For the larger games, I use very few different piece types, from 5 to 10ish, for ease of play, especially in multimove/activator games. I try to make games as simple as possible to showcase the idea of the game, for ease of learning and play, not necessarily for ease of winning.

A number of years ago I argued you could have a playable chess variant on a 100x100 board. Then I started looking for those games. As this topic demonstrates, it becomes very quickly realized that it has to be a multimove game, unless you want it to be a generational game, a family heirloom you pass down through the decades. I tried a number of things that didn't work (like Corps Chess or Spacewar.) The one idea that did work, and in more than one way, was using 'little kings' - leaders, brain units, all activators, to control small groups of pieces independently. The reason it worked is that large multimove games can get very chaotic very quickly, and leaders, along with using short ranged pieces, controlled this chaos, and actually used it to simulate the randomness of combat in large single battle games and even to simulate wars between 2 countries. So I've done successful 32x32 and 48x64 games, and demonstrated that the currently developed wargame version of activators could scale well beyond 100x100 and still give a playable, if long, game.

There have been 2 arguments against this idea presented here. The first is that the OP wants to use long ranged pieces, infinite sliders or leapers. Since this idea of combining leaders and long range pieces hasn't been examined at all, it is an open question, and I'll leave it there for a while. The second argument against military-style leadership is simple. If you have multiple moves, you can attack a single position multiple times in one turn, and that isn't fair. While I agree with the sentiment, I don't agree that multimove games have to be unfair. Regardless of how bad they may be as games, none of the fair number of activator games I've posted has been unfair in that way. The full wargames have terrain differences that can introduce elements of unfairness, but that is part of war. Further, the terrain is moveable. Change the board if you don't like it. ;) But activators are 'tunable'. The ones that activate pieces next to them (touching) or 2 squares away (and not touching) act like military leaders and supply units for short range pieces. (With infinite sliders, such a 'unit' would act like a moveable missile battery, an interesting military unit, no?) So, what happens if you tune your activator range to 'touching'?

Now you've got a whole different animal, so to speak, especially if you allow activation to flow through a chain of touching pieces. Now you've got something that can move and flow, that feels more alive. Since I'm naming-deficient, I call these 'pieces' chesimals, chess animals. Developing this idea for a very large game just might help out here. I will say this, chesimals need a lot of space, somewhat like real animals. They can't be jammed into all the squares of 40% of a large chess board and actually move - move at all, much less move effectively. SInce one "piece' is composed a several units, say 9 bishops, with the brain unit in the middle of a 3x3 square. Even when in the front row, the 3x3 square can't move without own capture if the chesimals are jammed side by side. To get out diagonally, the 'bishop chesimal' would need 2 files of 3 empty squares on either side to be able to move freely just in forward directions. And they tend to 'want' more than that. But you could set up each chesimal as a 3x3 critter in the middle of a 5x5 block of squares. So a 40x40 board section would hold 40 chesimals containing 360 individual pieces. Shoot, that almost sounds like it'd be playable in an afternoon! It certainly seems to beat the other ways I've been looking at a giant game. hmmm... You've posed a pretty puzzle here. Can you make a giant game using chess pieces and rules that remains a pure chess game playable in reasonable time, or must it involve elements of 'unfairness'?

General Discussion / Re: Should this Forum have an Off-Topic section?
« on: April 28, 2018, 01:24:28 am »
It's amazing how, and how often, that works! It's actually a fairly easy game, certainly the easiest and closest to standard Western chess 4D variant I know of. Grin, if you're ever interested, I'll cheerfully play a game, even if not very well. I've never claimed to be any good at it. In fact, it's about the only 4D vriant I can play and understand what I'm doing and even what my opponent may be doing.
Thanks for the offer. I might like to play sometime but unfortunatelly I'm a little too busy right now to add any more games to my roster. But one question I have - if we did play, at which website would you suggest playing?
I missed your question about websites. Sorry 'bout that! The only website I know of where you can play it is

Pages: 1 2 3 »