CCRL Discussion Board

Some considerations first : - 7-men won’t be available in the near-term future. Several years will be needed. - Final size depends of metrics used a lot. Current 6-men TBs (1,2 Terabytes) are in distance-to-mate (DTM) format, which is the most expensive in space storage. Current estimates for a full...

50 move rule is also currently ignored by composers, which makes full utility for them to use TBs ignoring 50-move rule, which are DTM, DTC, DTZ (and WDL). I find also that 50-move rule (if played with) raises very interesting theoretical aspects of endings such as kqpkq, knnkp, etc... in fact, the ...

I believe there is, in fact, a possibility (even if not thrilling) to build a DTM50 table in Nalimov format which also meets the requirement of giving an finite upper bound of the distance to mate from any given position, against any possible strategy from the losing side, considering only lines of ...

Your « old way » is in fact building WDL50 tables (which is already a project on 7-men !). These tables give you info about if a position is winnable under 50-move rule, assuming current movecount=0. It is by definition the win/draw/loss information which can be extracted DTZ50/DTC50/DTM50 tables (a...

- the set of positions marked as wins in the DTZ50 and DTM50 EGTs are the same, Absolutely right. One can even add that, for a given position, the DTM50 series indexed by movecount (as described above) has exactly 101-2*DTZ50 or 100-2*DTZ50 (depending on ply parity) finite values. - DTM50 recognise...

I’m with every syz’s points of view and explanations about DTM50. The main points about DTM50 are : - like DTC50, or “full DTR” as I name it, DTM50 doesn’t refer at (position, side to move) but at (position, side to move, movecount). Extracting DTM50 only for (position, side to move, movecount=0) in...

and with the move-counter in the index the tables most likely simply get too big I agree that with the move counter in the index (which is, as I currently believe, necessary for metrics such as DTR, DTM50 -or DTMx-, DTC50 -or DTCx-) tables will grow a lot, but their size will certainly be still man...

http://chess.jaet.org/cgi-bin/mzugs
Lots of info about mzugs here.
"Side to move loses" is full-point mutual zugzwang. Such hot positions are very few. There are none in kqrkqr.

Regards,
k

Irreversible/zeroing moves Really, I don’t think that making zeroing moves and irreversible moves match each other is a sound idea. That’s too theory oriented. Zeroing moves and 50 move rule are not intended to make chess match some kind of mathematics, but to make play enjoyable and, for what they ...

The indexing scheme used for generation will be built upon this notion. Otherwise the generator will not have the ability to generate the table slice by slice, which will lose a lot of performance. Also DTZ almost demands a slice by slice approach. Indexing scheme must be kept free, but if cutting ...

As a temporary solution to have something useful while waiting for 100TB hard drives, WDL seems a lot more interesting. (Another option would be 5-state WDL50+, if the difference in size turns out to be small enough.) Considering that : 1) WDL alone is far more useful than WDL50 alone 2) WDL50 will...

But I'm not sure how that helps much, as you'll need the subtables in DTZ50 if you want to use kpppkpp in DTZ50. Wrong. DTZ(50) needs only WDL(50) for daughter endings, not DTZ(50). This is the main strength of the “dual metric” approach : as soon as you have 7-men full set in WDL(50) format, you c...

Also the bounty may not become big enough to make it attractive when divided. Prize splitting was not a suggestion, it was just a possible future idea, should the overall project prove too hard or too long for a full solution to emerge. But for now I’m against this idea too, as I agree with your qu...

I suggest to use the following scheme for performance testing (this is also a proposed answer to some currently open questions in requirement sheet), upon which prize distribution rules may be elaborated, provided the quality of a solution, and not only its ability to make 7-men available in finite ...

I am curious what others think about this. I personally find the difference between 1. failing to win a won position because of the 50-move rule (as an obvious consequence of choosing the much smaller WDL over DTZ), and 2. getting mated after reaching a position on the board that is known to be a d...

In other words, DTZ50 will never blunder, and it will always make moves that look good to a casual observer or computer with similar disk space. Any tablebase that does better then DTZ50 would require an unfeasible amount of storage space. Take this example : Kc4, Rd1, Pd4 / Kd7, Bc7 d5 is the wors...

WDL or WDL50 tables can’t prevent KK’s scenario (engine starting from a won position but going in circles and finally conceding 50-move rule draw) from happening. This is due to lack of “winning line” info. This is the reason why DTZ or DTZ50 info is also needed for “not clear-cut” endings. On the o...

Some suggestions (these are not really new requirements, but details which might not be seen as self-evident when reading current requirement sheet). - Generator and probing code must be capable of 3-6 men as well as 7-men (including 51 and 51p). 7-men generation will use 6 men TBs as entry data, an...

Correct. But original question was : is it possible to map a DTZ50 table using only full-moves ? I don’t think so. True, distance to mate in plies is always odd (if winning side is to move) or even (if losing side is to move), but distance to next zeroing move is not. Let’s see an example : Ka8, Qc7...

I think koistinen is wrong as well. Since in a mated position, it will always be the losing side's turn, the superior side will always be moving on odd numbered plies and the losing side on even numbered plies. As long as you make sure the GUI is programmed properly, it should be able to translate ...

Good point to see that trying to minimize DTR as first priority can cause losing control of length of current phase. IMO, this points towards the same point I raised before : history of position (i.e number of non-zeroing moves previously played, let’s call it x) does have influence on DTR, and in a...

No, you're correct, current discussion about DTR aims at better understanding of this new concept, helping g (or syz) to program a DTR generator for simple endings (3-5 men perhaps) and trying to raise new knowledge with it on some interesting “target” endings such as kqpkq. This has nothing to do w...

***sorry g, I post this only a few minutes later than your previous post, and I haven't had time to take it into account. I hope this will be helpful, though.*** Syz, thanks for your support, but I also understand g’s arguments. Seems some misunderstanding comes from the definition of DTZR. If DTZR ...

sz's idea that one calculates a DTZ50 EGT, then relaxes the k-move rule by one to add the new 'DTZ51' wins and so on ... and that this gives DTR is quite wrong - on two counts. A suggestion : if I’m right, there is a straightforward and readily available path to DTR, provided a DTZk generator alrea...

KK noted somewhere that 'we will only note when the 50-move rule makes a difference for positions with DTZ > 50'. If I read that correctly, it is a big error. Please note that KBBKNN has maxDTZ < 50, and yet - because of KBBKN - a major % of the theoretical wins are subject to draw-claims under the...