H. G. Muller wrote on Sat, Nov 4, 2023 08:15 AM UTC:
There definitely should be quality standards on the presentation here, and it is important there is sufficient suport. If there is no way to play the variant on-line or against an AI, attracting attention to it becomes a pointless exercise.
But this need not be given much weight in the selection process, as it often can be provided easily within days after the selection is made. In this respect the selection process is badly flawed: rather than judging by the support situation as it could be when we make the selection, say, a week in advance, we judge by the situation as it is. This has the unfortunate side effect that it will also never lead to improvement of what we have, just to rejection. If there are 20 variants amongst the candidates that fail to meet the requirement of a computer opponent, there will be no incentive to provide one if there is no guarantee it will be featured.
For games like HectoChess, with no unorthodox rules other than the piece moves, computer opponents that play them can be created in a matter of minutes, with the aid of configurable variant engines like Fairy-Max, Sjaak II, Nebiyu or Fairy-Stockfish. If you use guestimated piece values (as ZoG would do anyway). Creating rule-enforcing GC presets is also only a matter of minutes for such variants, with the aid of the Play-Test Applet.
I think we should seriously consider providing downloads of configuration files for these stronger engines, rather than ZRF files that almost no one can use. That we don't do that for HectoChess is a missed opportunity.
There definitely should be quality standards on the presentation here, and it is important there is sufficient suport. If there is no way to play the variant on-line or against an AI, attracting attention to it becomes a pointless exercise.
But this need not be given much weight in the selection process, as it often can be provided easily within days after the selection is made. In this respect the selection process is badly flawed: rather than judging by the support situation as it could be when we make the selection, say, a week in advance, we judge by the situation as it is. This has the unfortunate side effect that it will also never lead to improvement of what we have, just to rejection. If there are 20 variants amongst the candidates that fail to meet the requirement of a computer opponent, there will be no incentive to provide one if there is no guarantee it will be featured.
For games like HectoChess, with no unorthodox rules other than the piece moves, computer opponents that play them can be created in a matter of minutes, with the aid of configurable variant engines like Fairy-Max, Sjaak II, Nebiyu or Fairy-Stockfish. If you use guestimated piece values (as ZoG would do anyway). Creating rule-enforcing GC presets is also only a matter of minutes for such variants, with the aid of the Play-Test Applet.
I think we should seriously consider providing downloads of configuration files for these stronger engines, rather than ZRF files that almost no one can use. That we don't do that for HectoChess is a missed opportunity.