CCRL Discussion Board

If the difference is just 1% in both space and access time, for both DTZ50+ and WDL50+, then I don't have any problems with DTZ50+ and WDL50+. Note that it is WDL50 (or WDL50+) that needs fastest possible access, not DTZ50+. Note that going for WDL50+ instead of WDL50 will need capability to store ...

Is "DOS" the same with what I called WDL(>50)? I don't like the name "DOS" because DOS is used for many other things already. May be DUS, then ? U = Unaffected position by 50-move rule. Or anything else. IMHO, it’s up to the community to find suitable terminology, if no one alre...

Can a correct en-passant-aware tablebase be generated without considering en-passant rights in root? (Or without storing the positions with the en-passant right). I don't think so. Please correct me if I am wrong. Ah yes. You're correct on both. I missed second argument, positions with en-passant r...

kronsteen wrote: All Tablebases make the assumption that evaluated positions contain no opportunity for en-passant pawn capture for any side, even if pawn position is compatible (but en-passant rule naturally must be considered in subsequent play) Here I have to disagree. En-passant is a very impor...

According to what has been already discussed, I would like to submit some first-hand text in order to help starting building a project. Maybe it would also be the opportunity (if not too ambitious) to try to raise "standard" terminology about won positions forced into draw by 50-move rule,...

In chess studies, the 50 move rule is not used. This is why the proposed DTZ50 is not useful for me, or to the other chess composers. DTZ50+WDL50 tables will fully meet chess composers' needs, provided tables hold info about cursed positions, as I suggest. A cursed win is a win for problemers and a...

A fourth position state will affect final size, unless one uses the very naive compression of packing 4 positions in a byte. Such compression would give a compression ratio of 25% for a WDL-table like KQQKR (starting from 1 byte/pos), which is extremely bad. (Slightly better is packing 5 positions ...

DTR/DTZR : To me, DTR/DTZR metrics are best aimed at two goals : - building playing programs which perform well against humans in endgames - providing new knowledge for theorists which want to process it in order to extract useful practical information for players, and write books for example. For s...

DTZR = the least number of moves to complete the current phase, given that White/Black are minimaxing DTR as 1st priority. Should that not be "the" number of moves to complete the current phase? What ends a phase for you? A zeroing move, a move that lowers DTR, or either? (My guess is tha...

I agree with g, DTZR solves the problem of DTR I mentioned above, as it shows winning lines. It is also clear way of progress to show, not only winning lines but also “good ones”. It opens a new path for TBs, moving from purely theoretical knowledge to practical knowledge where purpose is to maximiz...

To follow on discussion about metric choice : DTZ/DTZ50 The more I think about DTZ50, the more I like it. Space saving and relevance to real chess sound like big arguments. DTZ has the advantage of being immediately connectable to existing 6-men in DTM format (which is not the case of DTZ50) but it ...

I you think of DTZ, I may think that DTZ50 (i.e TBs taking into account 50-move rule) may be even a better idea, as it saves even more space and rises a new very interesting aspect of the game. I don't know what work has currently be achieved about DTZ50 (5 men ? Some 6 men ? Specific interface for ...

Greetings, I’m a French chess amateur, and I enjoy tablebases a lot. Here are my opinions about your proposals : Metrics : Code should be able to provide some kind of DTx. I find DTM more fun and more in the spirit of the game, but maybe space savings by using DTC or DTZ are too much important consi...