Check out Symmetric Chess, our featured variant for March, 2024.


[ 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

Earlier Reverse Order Later
PieceTypeLists[Subject Thread] [Add Response]
George Duke wrote on Thu, Dec 8, 2016 08:17 PM UTC:

Recently Knappen found from index of Jelliss Ski-Rook type, to compare to pieces in current Chess Different Armies Asymmetrical Chess. There are twenty or more long glossaries of piece-types from Charles Gilman's to David Howe's. So let's put many lists of types here in new topic saving going back to locate them. This French one came to our attention eight years ago: Christian_Poisson.

George Jeliss' best list for types: Fairy.

And Knappen cites German one then too: Types.

M. Winther's own Bifurcation pieces, inventing and adding to renaming about six Jeliss had earlier recorded: Bifurcators.


George Duke wrote on Tue, Dec 13, 2016 07:20 PM UTC:

"Analyze the motion of a smooth flat coin rolling inside the rough surface of a hollow ellipsoid balanced on the back of a hemispherical tortoise ambling at constant speed straight up a hill of uniform gradient on Saturn, of Sol." In that problem many assumed values can work the math with the right equations.

Likewise piece values, changing with board and array, are problematical compared to mere definitions of types. This topic is for Lists of piece-types, so that all of them are accessible. The first samples were one French Glossary, one German, one by George Jelliss, and one by M. Winther. Notice Winther estimates value for each defined piece. To add about 20 more long lists, Charles Gilman is very accurate once you understand what he means, and he never does values. In Index just under 'A' there are about 150 definitions: Letter_A.

Continuation to 'B' and through the alphabet turns up around 2000 piece-types -- without values. Gilman and Betza are the only designers who saw fit to make over 150 CVs. In contrast to Gilman, Betza of course is full of estimated values. Betza was always the best at including values and also at giving sample game scores. Betza never did a glossary though. (And for that matter he never went outside the simple 8x8 box, except Outrigger 8x10 and his particular Double Chess. Gilman has every imaginable board size and the most three-dimension CVs ever done.)

Half of 'S' by Gilman:

S_1/2 End of each leads in next letter.


George Duke wrote on Wed, Dec 21, 2016 09:29 PM UTC:

Next for piece definitions in long lists is Truelove's table:

Pritchard, compiled from Pritchard's 'Enclyclopedia' and Dickens 'Guide'. If not recognizing the name immediately, you have to go to the book to see how the piece-type moves, or it might be found in a CV game online or in another glossary. Though not trying values, Truelove's is useful because some games in Pritchard have not made it over to Chess Variant Page.

Not useful from the same turn of millennium time period is Derzhanski's:

Fairy The point values (zillions) are presumptuous in exactitude, and supposed definitions, whether accurate Betzan notation or not, are like personal notes, as if Derzhanski didn't know what he was getting into.


George Duke wrote on Thu, Dec 29, 2016 08:14 PM UTC:

Next, this is the piece section of David Howe's "Concise Guide" put out in 2011: Pieces.

And not far afield Howe's list of categories of types from earlier 2001: Types.

And then Christine Bagley-Jones' historic 150+ Shogi variant pieces found in this 2012 article: Fairy_Shogi All 154 piece-types exactly are in fact indicated in the one 12x16 first board of "Fairy Pieces." Frivolously, if you make the board 24x16, there can be two sides with 154 different-moving pieces Shogi-derived from actual recorded Japanese CVs, but of course most of them are unable to move from respective array.

It was found that Gilman's combined indexes define about 2000 differing piece-types, only 75% or so newly made up by him. Thus to show all Charles' the way Christine does diagrammatically on just one board can easily by done for the Gilman 2000 on a given 70x70 chessboard, providing for both sides. Again, most would be immobilized at the start until things open up by areas of board size 70 square. [By coincidence -- more precise than rounding to tens -- 64x64 board gives 4096 squares and that is the first size more than enough too if omitting Pawns, to exhibit often-repeated 2000 piece-types each force, plus a King and some space between.]


George Duke wrote on Sat, Dec 31, 2016 06:27 PM UTC:

Shogi_Fairy. Christine Bagley-Jones' pieces on the 12x16 board, "Fairy Pieces," are from the historical large Shogi variants pre-19th century. Some more could have been added she explains, but displayed on board in the article are all of Angry Boar to Yaksha for 154 types.  Yaksha pictorially is on j10, Coiled Serpent e2, Dark Spirit h2, Strutting Crow d9, Bishop e1, Knight ("el")-1, Rook b8 and so on. Since no one has marked the squares yet by description of the piece, the way to find them is from the definition to the board not vice versa.

Charles Gilman's 2000 piece-types -- his entire work -- could be exhibited together on one board too. To show them all as a CV to play in set array, mirrored for both sides, a chessboard of 64-square suffices. A logical CV with one of each of the 2000 different pieces would dispense with Pawns and add one King, say at ff1 for White and ff64 for Black.  Since 64x64 is 4096, initially the pieces will cover the entire board but for one empty band of squares staggered symmetrically from a32 to lll("els")33. There are then 94 vacant squares between the two armies to start, and no opponents abut each other. This CV, call it Gilman-Complete, might best be played Marsellais-style or even Progressively.

Go boards are 19x19 and Scrabble boards 15x15. Three Scrabble boards plus one Go board make 64 squares in one row -- ignoring that 18 squares of 19 intersections usually make Go board. Just as suggestion, there would be different ways to size and space quickly to make playable surface of 64x64. One far-fetched way for visualizing anyway is 9 scrabble boards formed into larger square (45x45) and 6 or 7 Go boards surrounding, sketching in the needed extras. That's what it takes to include all Gilman's indexed types. Or four of Bagley-Jones' 16-wide boards above give enough files 64, and build up from there to get equivalent of 64 regular chessboards. Scrabble and Go_boards.

Charles has four geometries chiefly: cubic, square, hex-prism, hexagonal (excluding cubic xyrixa of Tetrahedral). Gilman p-ts are from all four, so to be fully playable, there needs to be agreement on rules for transforming 3d to 2d and hex to square. In general make the move first by its geometry, then translate if necessary to the planar square-based board of 4096 cells. Mixed geometry questions like this can be found in Betza's Rectahex, Geometry, its comments, and several Gilman CVs themselves. Cubic pieces and hex-prism could readily be resolved to play one of each type on flat 64x64 for the long haul.


George Duke wrote on Sat, Jan 7, 2017 05:57 PM UTC:

In David Howe's Megachess every piece is a chess game: 32_boards. If putting a unique piece on each square of 32 boards like Megachess has, that suffices to see all of Charles Gilman's approximately 2000 different piece-types and leaves 48 vacant squares for additions and omissions. "The Mega-board (detailed view)" of above article, including the empty squares marked, has all the space needed to number the Gilman types (of course not done in Megachess).

Charles' recording is actually 21 separate indexes, for example 'W' includes W, X, Y, and Z. The 'I' section: I_pieces. The "I's" include 'J' and so "Japan" and there distinctly lists most all of those in Christine Bagley-Jones' article on Shogi Fairy over again. Then one has to look them up for how they move under the Gilman-approved name.

For example, Bagley-Jones: (name Phoenix) (help "Phoenix (Hoo): moves one square orthogonally or jumps to the 2nd square diagonally.") (description "Phoenix\\A Phoenix may move one square orthogonally or jump to the second square in any diagonal direction.") And then Gilman under 'I' above and to "Japan" notes "Phoenix(Hoo) Waffle" and then to "Waffle": WAFFLE square Wazir+Elephant ¢.

Next, Jeremy Good has a short list that's a good refresher: Complementarity.

And Betza's Bent_Riders. That makes 12 separate-authored piece lists. Gilman's are organized into the 19 first letters, with C and S further split into two.


George Duke wrote on Mon, Jan 9, 2017 07:53 PM UTC:

Then there are Betza's "Ideal and Practical Values" that also list pieces. The basic or atomic chess pieces: Wazir+. I don't like Ralph's use of "fundamental" for W, A, D, N, F, because fundamental should be reserved for Rook, Knight, Bishop, and Falcon; but those five FWAND are certainly the atomic ones for 5x5 around a starting square.

Just as Winther has "Bifurcators" in earlier list here, Howe has "Mimics": Mimics. There is much unrealized potential in having CVs with Mimic/Mime/Imitator.

Next, I list pieces that are riders with more than one path: Multi-path

Joe Joyce's and Christine Bagley-Jones' Short-Range_Project.

(Negative-value pieces fit this topic as well, Nattering_Nabobs, because Negative Relay, as one way to achieve over-all "minus value," could be generalized as Mutator to cover the 2000 Gilman types and approximately 500 other types in these lists so far Gilman happens to exclude or overlook. Just operate the Mutator not on the game-rules per se but the individual piece-type directly. Presto, 2500 different piece-types become 5000.)


George Duke wrote on Tue, Jan 10, 2017 05:53 PM UTC:

This thread is to place the important piece lists together. Of course there is some or considerable overlap of piece-types, such as Charles' Index under 'J' including all of Christine's Japanese Fairy Shogi pieces from 12th to 18th centuries.

Next there are lists or discussions embedded into CVs themselves. Fergus Duniho's Gross_Chess covers basic Chess pieces, the most popular compounds from 17th century Carrera on, and hoppers all in one essay.

Fantasy Grand, F_Grand, in its six different armies has dozen or more pieces, as compounds or limited movers, not found elsewhere, serving as plausible piece-type glossary. Notice that in the Giant Army the Cyclops is the Ski Rook. Jorg Knappen recently told Dmitry Eskin that his current Asymmetric Chess is using longstanding Ski Rook. Thus Ski Rook is piece-type in two Chess Different Armies: Fantasy and Asymmetric. Eskin must have borrowed the idea from Hatch.

In fact, the first comment here starting this whole topic was that George Jeliss does define Ski Rook: Ski_Rook. What goes around comes around.

For follow-up, Suffix, Gilman develops the Suffix Index which can multiply the 2500+ piece-types formulaically.


George Duke wrote on Thu, Jan 12, 2017 08:04 PM UTC:

In only eight comments here are Lists of specific piece-types clearly defined, one by one, numbering about 2500. 2500 different pieces not enough to choose from?

Suffix Use Charles Gilman's Suffix Index to generate multiples more.

Take the first one ALTERNATOR:

ALTERNATOR after angle (rounded to nearest degree on these pages) and full name of 2 pieces square hex a piece whose move's odd stages are those of the first, and even ones those of the second, piece, each in one direction. An alternator can be prefixed MIRROR to start with the second-named piece, CONTRA- to end with the first-named piece (and start with either), MIRROR CONTRA- to end with the second-named piece (and start with either), and DOUBLE that to be free to start or end with either.

Say you want to combine Wazir and Knight, not as compound but as combination or sequential piece. The first new type of piece is "27 degrees Wazir Knight," meaning a piece that moves first as W then N in the narrow direction as one move. 'Mirror 27 W N' starts with Knight then Wazir, and in a class you may want to keep the naming order the same and so actually use "Mirror." "Contra 27 WN" must stop with a Wazir but can start with either. "Mirror Contra (implied 27) WN" has to stop with a Knight. "Double WN" is higher value because can start or end with either. There are other possibilities including making the angled change of direction 63 degrees for the wide Knight mode. And there is the opportunity to allow both 27- and 63-degree angled changes of direction. Estimate there are about 20 reasonable p-ts by combining Wazir and Knight in different ways as ALTERNATOR.

WEAVER is the last suffix alphabetically. A weaver is an alternator by definition. If we make all the above special-case Weaver instead of plain Alternator, it means the Knight's direction changes systematically every other step in its leg of a solitary move -- that is alternating between Wazir and Knight as well as weaving like a crooked nightrider. Jorg Knappen finds in "Nachtmahr" there are quite a few Nightriders so it needs specification between which two the Knight weaves as part of the ALTERNATOR. That is just a matter of putting two different angles in the piece description, or else using accepted name for it from Knappen.

Probably 100 reasonable piece-types can come about from combination piece -- meaning more than one leg -- of Wazir and Knight. Gilman's system can describe the movements well enough, and it's not worth trying to adapt simple Betzan notation become anachronistic.


George Duke wrote on Thu, Jan 19, 2017 08:43 PM UTC:

You can Name the Baby anything you want but over half name it by these 100 in both lists: Top_100. Likewise variant pieces draw from about 100 choices more than half the time (popularity in the eye of the beholder). For convenience in comparison, educated guesswork can deliver specific match-ups: that #1 is Mary/Centaur(BN) about same-percentaged, #2 is John/Patricia/Camel, #3 Jennifer/Alfil, #5 William/Linda/Commoner, #10 Sarah/Cannon, #15 Donald/Champion(RN), #100 Howard/Kathy/Padwar of Jetan, and so on in between.

If we liberally cut piece-types off at 2000 and CVs at about 9000 from all sources, including some subvariants, and assuming about one new type (not rnbkqp) per CV, it means 2%, or 180, CVs use Centaur/Sarah and 90 CVs use Camel/Patricia. Centaur and Camel may be low a bit from more expected 3% and 2% respectively, but most others being employed near 1%, or less, of the time make for good comparability between Babies and Fairy pieces. Remember half of CVs are on small 8x8 where #1 Centaur(BN) appears only rarely, and half again of the 64-square ones may have zero additional pieces. And more pertinently, there are many important specified pieces used only once or twice -- once per 9000 is about 0.01%. Most types will fall in the range 0.009% to 0.7% usage.

Beauty and eventual popularity can come about by developing formulaically: just see the 'Man & Beasts' project. For example, try compound of more frequently-used p-t and infrequently-used p-t. Noting "Howard" is #100 on the Boy's side, and William #5, there is (Commoner plus Padwar -- Padwar moves two steps diagonally allowing the bend). The compound is neat brand new type, and is fitting because William Howard Taft is USA President and Chief Justice. Taft is the ...er heaviest, well fattest USA Pres and Trump number three after Cleveland. In addition to the oldest ever Trump is third in ...weight. What is Donald John Trump's piece-type? See above that is (Camel plus Champion(RN)) already named "Carshall" (or Car-shell? or Car Shill?) by Gilman: Marshall+Camel.

Name_Game.

Alphanumerically, trump and truman are only off by one. Trump's butterfly in hope he'll do about-faces on Nature.


George Duke wrote on Sat, Jan 21, 2017 04:47 PM UTC:

There are 9900 (100x99) variant chess pieces attained by making bi-compounds from the Name List. Two of them: William Howard Taft is Padwar plus Man, and Donald John Trump is Champion(RN) plus Camel (see last comment here). It requires assigning a piece-type according to prevalence to each of the first 100 elements in the population. Since fortuitously #100 in both listings occurs around 0.2% of the time, there is surprisingly close correspondence.

Taft, Padwar & Man, golfing: Outing; and Trump, Cam-shall golfing 100 years later: Round. Round2.

Ca-shall -- or Car-Shill or CarShell, they're all okay -- is by the Gilman nomenclature for Camel compounded of Marshall, with Marshall(RN also) being one popular Champion re-naming. In other areas of Gilman's system there is stricter distinguishment the exact form of the name, but with "Marshall" as one leg already a compound itself, there is flexibility since Gilman gives method of starting Camel compound with "Ca" but not precise name this or every case, where there could be diagram and word definition too.

Gilman Suffix_Index, as well can enhance pieces derived this way from the serviceable Name List. If using "-Lander," there is Ca-shall-Lander, or TrumpLander as one will, which piece moves like Camel or Knight but can only step one square orthogonally, not permitting second space and beyond of the full Rook mode. That is synonymous with (Wazir + N + Camel), and designers may prefer explanation by one form or the other according to context, such as what other pieces are being developed. Mathematical equivalence.


George Duke wrote on Sun, Jan 22, 2017 08:24 PM UTC:

A portmanteau using Gilman methodology for the compound of Champion(RN) and Camel is Carshall coming from Camel and Marshall. By application of the Name List, it can be called Trump too. Carshall, Trump, Car Shill etc. are one and same compound (Rook + Knight + Camel). There is liberty with the name since Charles did not do this series completely.

Then bringing together the Suffix_Index and Name List, there are by extension untold millions of personalized bi-compounds. For example, Trump-Charger is much weaker piece than Trump-Lander, the latter it was seen being omni-directional. All Chargers are Landers by definition, with further restriction of only going forwards, uni-directionally now. Depicting piece as Charger -- already single-stepping or single-leaping by virtue of being Lander -- makes it perfectly Pawn-like. This particular type Trump-Charger then can move forward as Knight, Camel or Wazir, a great five choices for a Pawn. There is no provision for Promotion in the definitional naming itself, but it would be logical as added Rule of the Game.

Let a Pawn be both Berolina and OrthoPawn at the same time, more than doubling its value, and it is accurately described as "Queen-Charger," explaining separately the initial two-step option if wanted.

Differently, how about TrumpRanker? Trump is unambiguous piece though having synonymous names CarShill etc., and "-Ranker" just requires moving more ranks than files. Thus Trump-Ranker is restricted to the narrow Knight and narrow Camel moves forward or backward and only rookwise rankwise (for any distance).

Next, Trump-Filer moves by wide Knight, wide Camel, or Rookwise filewise any distance. Then, Trump-Blinker (or CarShallBlinker etc.) moves as Trump-Ranker but captures as Trump-Filer -- a divergent piece like the F.I.D.E. simple OrthoPawn. CarShill-ContraBlinker, same as Trump-ContraBlinker, moves as Filer any of the compounds three ways (N, R, Ca) widely in the technical sense and captures as Ranker by the same three narrowly.

In another category, non-divergent TrumpRover can indefinitely repeat the Knight legs or the Camel legs, but can only step as Wazir orthogonally. Its purer Gilman name is Carshall-Rover. Still another category by entering another Suffix, '-Switcher' requires odd number of times in one move, so Trump-Switcher goes any of the 20 directions (4+8+8) by stops on the first or third or fifth and so on if board is big enough, whether as Rook, or Knight, or Camel (but not combination) -- nice piece for 16x16 and up.

Yet another group, if ending with '-Snatcher', like Trump-Snatcher (CarShall-Snatcher), the piece-type can repeat a capture in the same direction on the same move (presumably only the very next step or leap). Combining the last two, a Trump-Switcher-Snatcher goes the 20 directions, stopping on odd legs and permitting immediate repeat capture(s) by direct continuation.


George Duke wrote on Mon, Jan 23, 2017 05:58 PM UTC:
"Don't you think I could also be a Grandmaster?" asked Donald John Trump. Benko. No, Trump can only be commemorated in a variant piece, as here.

PUSA_Chess Warren Harding is one more definite Yes contrary to left article. Board.

R+N+Camel.

Okay it turns out Gilman does name Champion(RN) plus Camel! Can you find it in 'M&B8' there before looking at the follow-up comment?


George Duke wrote on Mon, Jan 23, 2017 05:58 PM UTC:

ACROPOLIS. "For a pair of duals or semi-duals as oblique components I start with Gnu compounds, using the reversed letter pair of the Ace/Acme group. Rook+Gnu=ACROPOLIS, after a lofty stronghold. Bishop+Gnu=ACTOR, which like MAB 06's Clown and Buffoon draws on the Bishop as Fool with a performer. It is worth noting that surnames such as Abbot, Bishop, Duke, King, Pope, and Prince originate in actors typecast in such parts. Queen+Gnu=ACTRESS mirrors the femininity of its radial component." --'M&B8'

Above Charles Gilman has in fact named (Champion(RN) + Camel). ACROPOLIS by Gilman is the same (R+N+Camel) we are calling Trump. From now on TRUMP or ACROPOLIS will do for the piece-type here, Charles would agree, and we can drop the experimental portmanteaus CarShall, CarShill, CamShell etc. (I thought I had seen it but unsure what form.) There is nothing wrong with an occasional "Carshall-Rover" and so on, using the first letters of Camel for clarity since they are perfectly understood, but better is ACROPOLIS-ROVER or TRUMP-ROVER where there is precedence. One good alternate name is enough, when more can be avoided. Charles does approach it from longstanding Gnu(Knight + Camel). (Q+Camel) was already named ACME (and (Q+N) was re-named ACE by Gilman, but most will keep AMAZON). It all shows both the flux and the systemization in nomenclature.

.

14 comments displayed

Earlier Reverse Order Later

Permalink to the exact comments currently displayed.