| |
BGonline.org Forums
Nactation after the initial moves
Posted By: Nack Ballard In Response To: Nactation after the initial moves (Andreas)
Date: Thursday, 26 November 2009, at 1:05 p.m.
Andreas: Nactation is an exciting approach for a new language describing BG moves with intuitive plausible shortcuts. Very successfull obviously at least for the reasonable 1st and 2nd moves so far. It seems attractive to expand this language deeper in the game.
Nack: Nactation caters more to describing the first few moves of the game, but as long as non-doublets are rolled (or doublets are played in pairs) it should be easy to nactate middle game and endgame moves.
Andreas: Nack has recently demonstrated here an example how Nactation can be used for describing the moves until deeply into the midgame. Nevertheless it seems often not easy to find the proper Nactations with the so far established characters.
Nack: It depends on what you mean by “so far established.” I’ve explained to a few enthusiasts (mostly in e-mails) resources that do not yet appear in the tutorial -- some of which will be shared in this post, thanks to you ;) Also, the ability to “find” the right character is often enhanced by practicing what you do know.
Andreas: How would you e.g. nactate the possible moves after rolling 43 in the following position after 63R-42P:
Andreas: 15/8 -> T? V?
Matt: I prefer L. T is ambiguous with 13/6. V just doesn't look right. I'm not sure if there is a stated difference between L and l yet.
Stick: T would be my choice, you could also use 8. (I'm sure Nack will tell you himself how to properly nactate everything you asked but I'll give you my quick take on them, as I would nactate them OtB.)
Nack: There are three answers for 15/8. I use L (Lift, i.e., moving a blot to safety) as Matt chose, because in most cases I prefer using letters (to contrast with the roll, which is in numerals), but Stick’s 8 is perfectly fine. [Matt, “l” (lower-case L) is used for a lifting play with a destination that is further from the 6pt than that of the L play, when there are two plays that fit the term “lift.”]
T (sTack or Tower) is technically wrong here because it means 13/6. When there is a choice of points on which to stack, the one closer to the 6pt gets the capital letter. Lower case t works for 15/8; that is, T = 13/6 and t = 15/8. Closest-to-6pt is a convention that can be applied to most letters. (For equidistance from the 6pt, tie goes to the lower-numbered point).
In other words, the definitions overlap for 15/8: it is both a blot-lifting and a stacking play. By contrast, 13/6 stacks but it doesn’t lift. If you feel unsure of the distinction, refer to Section 5 of the Nactation tutorial.
[I agree with Stick’s implication that when nactating matches for yourself, you can afford to be more liberal with your choice of characters. You do what you can in the heat of the moment. However, I would hope, ultimately, to see a standardized usage in communication with others.]
V is not such a broadly applied letter, Andreas. V really means “Variant of Up,” not just “Variant.” As a memory aid, V physically resembles U and follows it in the alphabet. (Refer to U in Section 2 and especially V in Section 5.) V is U’s back-up (just as d is D’s back-up). The reason I adopted a second letter is that there are sometimes more than two up plays: the expanded hierarchy for up plays is U, V, u, v.
Andreas: 15/11 13/10 -> V with a "down default"? D with an "escape default"
Matt: I would use d.
Stick: D would be my choice by assumptive Nactation I know there's no reason to leave the blot on the 15pt so all I'm nactating is the other half of the play.
Nack: I’m strong on d here. That is, 13/9 13/10 = D and 15/11 13/10 = d. If Matt got this right by applying the 6pt convention (i.e., 13/9 (D) is closer to the 6pt than 15/11 (d)) – which I don’t remember explaining to him(!), then kudos.
[Nactation hierarchies balance two aspects. One aspect is that the more prominent character (e.g., capital letter) describes the better play more often than the less prominent character (e.g., lower-case letter) does. For example, after a deuce-split and subsequent roll of 63, 22/13 (R) is usually better than 24/15 (r).
The other aspect is that rules for hierarchies should apply as universally as possible. Sometimes the two aspects are in conflict, as is the case with D/d here. Moving a 15pt blot down is usually better than moving a midpoint spare down, but I decided to arrange it so that the midpoint-down gets the capital letter because otherwise you would have to remember that D/d (or perhaps any move involving the outer board -- whatever is toggled) is an exception to the closest-to-6pt convention.]
Employing assumption (assumptive Nactation) for the “obvious” 15/11 part of the move may work for Stick, but to me that seems too casual for a general audience. One problem is where to draw the line when 15/11 is obvious and when it isn’t. For example, with 63R-54X-41, Black enters hitting with the ace and plays the 4 down. But which 4 is D and which is d? 13/9 and 15/11 are equally strong. The 6pt convention avoids fuzziness. I tend to use assumption only when obviousness overwhelms and/or the situation conforms to no existing convention and/or a superior choice of character eludes me.
Andreas: 13/6 -> D?
Matt: 6 would be my choice. Maybe something like i is OK (I meaning the really atrocious 8/5 8/4). T or l is probably OK but everyone will think these refer to 15/8.
Stick: I'd use 6.
Nack: Either T or 6 is fine. I’d use T for an audience familiar with the 6pt convention because I like using letters; however, 6 is always safe, and I'd be especially tempted to use 6 when it refers to as inferior a stacking play as 13/6 is here.
I haven’t yet created a strict definition for “I” (In) but it tends to be for moving a single checker in on the somewhat rare occasions I use it. Currently, I is okay for 13/6, and I suppose i would be safer, though in truth I’m not enamored with either.
For 8/5 8/4, I use A (Attack/Aggress) or & (double slot). I seem to recall that Herb Gurland once got to play a hilariously positive equity prop against opening 43A.
Andreas: 24/20 13/10? -> B? U with "down default"?
Matt: I'd use B, seems clear to me.
Stick: Here I'd use S as it's your standard big split play just in an unusual position.
Nack: Like Stick, I prefer S. The comprehensive definition of S, for the “splitting” half of the move, allows playing up even when the back checkers are already split or there is only one of them. The other half of the move comes down, as usual.
I rarely use B for non-doublets, because I haven’t “needed” it, and I haven’t decided whether B should conform to S’s larger number splits (which sounds natural but has the disadvantage of being in opposition to the closest-to-6pt convention) or whether B should be treated like Z (reverse split). [For doublets, B moves the same-sized number in both places; in that limited sense there is no reason to prefer S over Z, or Z over S.] If I decide B conforms to S, then 24/20 13/10 is B (chosen by Matt and Andreas), 24/21 13/9 is b, and (with the rare italics bird rearing its head in the hierarchy) 24/21 15/11 is B.
Andreas’ second suggestion of U (with down default for the other half of the move) is passable though unorthodox.
Andreas: 15/11 8/5 -> $ ?
Matt : I'd use $.
Stick: I'd use $ but it feels technically wrong. Again, I'm recording the match and seeing all the moves, a move like this by my opponent I won't forget so it serves its purpose even if I know it's the wrong symbol.
Nack: As Stick suspected, $ is wrong. The right answer is % or o. It’s your fault, Andreas; where do you come up with these plays? :)
In Nactation, the primary definition of slot is to move a checker down and put it or another checker onto a vacant offensive point. To feature the most “natural” slotting moves (without too many rules), the priorities for slotting hierarchy are: (1) slot the 5pt or 4pt, (2) directly unstack the tallest point possible, and (3) apply the 6pt convention.
Three slotting plays equally satisfy (1): 13/9 8/5, 15/11 8/5, and 13/10 8/4. All are equal in regards to (2). So, we go to the final tie-breaker of (3) the 6pt convention. Playing to the 9pt and 5pt is awarded $ (the gold medal symbol). Playing to the (more distant) 11pt and 5pt is awarded % (silver medal). 13/10 8/4 (which doesn’t put a checker as close to the 6pt) is awarded italic $ (bronze medal).
As a memory aid, % (alternate slot) is to the right of $ on the keyboard. It is also possible to perceive the two little circles as the top and bottom of a capital S and the (slanted) line though it as a lower case L (vaguely in the same way that $ is an S with a lower case L through it).
An easier letter (set) to implement is O, o, O, o. You apply only the 6pt convention, and the hierarchic order of the three plays is the same: 13/9 8/5 = O, 15/11 8/5 = o, and 13/10 8/4 = O. The letter O stands for “Outer” and means half the play is made into the player’s Outer board and half is made out of his Outer board. See “O” under Sections 3 and 4.
Andreas: Has there already been tried to nactate an entire game or even a match ?
Stick: When I record my matches I nactate them, traditional notation pains me. While I'm sure there are some technical mistakes in my on the fly Nactation, it can get the job done and I understand everything I meant at the time.
Tim: Implicitly I think you're asking whether Nactation is powerful enough to handle every possible situation. Nack has been kind enough to include me on some email discussions where he has revealed that he is still continuing to develop the system and add more symbols for rare situations. One can infer that there are still some tough cases that Nactation, at its current stage of development, is not quite ready to handle. I imagine one could dream up a not-totally-implausible position where there are more ways to play double 1's or double 2's than there are distinct Nactation symbols.
However, I'm sure that there are many games and even matches where you could use Nactation for almost all, if not all, the moves.
Stein: I have tried a little. Opening and middle game seems to work very well - basically whenever different plays follow different plans, there usually is one or more ways to nactate each play.
Where it gets difficult is where all plays follow the same plan, with slight differences. Typically, in a race bear-in, you may well have a dozen plays that tries to bear in as flexibly as possible, where the only real difference is exatly which points you put the checkers on. But, as Nack himself says, nactation isn't all or nothing, you can always fall back on long notation, whenever you feel that no nactation letter is good enough to describe the play.
Nack: At a tournament many years ago, I found a pocket of the playing room with only a few spectators. I positioned myself between two rows of tables where with minimal roaming I could view six matches (two of which were pretty slow-paced). I kept up with all of them. I stayed for two or three full-length games, long enough to convince myself that simul-transcription was not overly difficult. (It wouldn’t surprise me if Stick Machine can do the same.)
If memory serves, I used two characters instead of one on about 15% of the plays, in some cases because I wasn’t quick enough (trained enough?) to think of a workable single character, but in more of those cases it was because the system had not evolved to the point it is now.
The development of Nactation is on a need-to-use basis. I use it for the study and compilation of early game data. I haven’t transcribed matches in a long time. However, if people show enthusiasm for nactating games/matches and present moves that were actually played or might well have been played (as opposed to positions fabricated to challenge the outer envelope of the system), I’ll gladly answer any questions.
Tim supplied a perceptive answer to Andreas’ final question. Having applied Nactation almost exclusively to the early game for the last few years, I don’t know if there are plays that might reasonably arise later in the game that I can’t yet nactate unambiguously. I’d guess there are. Reading some into Stein's observations, for example, it is quite possible that certain bear-in doublet plays are yet to be solved. But I don’t feel motivated to expend the effort creating further nuances to peg the remaining plays unless some general desire for that level of perfection exists.
[Btw, I believe that single-character Nactation could ultimately handle any play, even for the wiliest compositions designed to confound. But to get to that (impractical) point would require even more work (probably a lot more) than to conquer 99.9+% of plays that might arise in the course of normal backgammon. Consider, for example, that every character could be given a strategically-designed secondary definition that is available when there is no legal play with the primary definition, and a tertiary definition behind the secondary one. (That should be plenty even without tapping into new alphabets!) In short, it’s not a question of whether any conceivable legal play could be written with a single character, but rather why try.]
Thank you, Andreas, Stick, Tim, Matt and Stein (names listed in order of posting) for your interest and participation.
Nack
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.