| Author | Message | |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 12 Registered: 5-2016 |
Gyges is an abstract boardgame with very few, very simple rules, of the kind ZoG usually excels at. See here:www.gyges.com. Claude Leroy, who invented it in 1980, kindly gave me permission to try and make a ZRF out of his concept. ZoG mostly seems to correctly handle the two main rules, but will allow the human player -that's me, to play by one rule only, even when one tries to reproduce the moves performed by the engine during an auto-played game. That effectively prevents me from trying to further debug the ZRF, which is obviously wrong in some aspect(s). Here it is for your kind scrutiny. I would be very grateful if someone could pinpoint the faulty parts. Thanks for any help.
| |||
| Sean Duggan (Dream)
New member Username: Dream Post Number: 120 Registered: 9-2000 |
Can you clarify which moves it is that the human can't play? | |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 13 Registered: 5-2016 |
Try making a displacing move, you will find ZoG won't let you and will restrict you to bouncing ones only. Now in the ZRF, if you put the displacing moves (named Bully) before the bouncing ones (Snooker), the situation is inverted. Pretty weird, considering both are assigned equal priority as they belong to the same move-type, namely "GoingOn". Beats me. In the (demo-)plays I provided, try repeating the moves labeled (zug x) in the ZSG file names to see the fact for yourself. | |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 14 Registered: 5-2016 |
I rewrote the ZRF to include attributes. It presented the same flaws, but as a pis aller I adapted it to a proprietary pieces variant. That spoils the original idea, but the result is rather fun in its own right, if one overlooks the tendency of the engine to often tirelessly run in circles (ZoG does not consider repeated series of partial moves a repetition). Also, though I never witnessed it done in this variant, a piece may get stuck; ZoG will continue playing nonetheless. Maddening, but here goes:
Mind you, Zog is pretty easy to beat. | |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 15 Registered: 5-2016 |
Well, I eventually did encounter cases where a piece gets jammed. It struck me that I had forgotten to state the "pass turn" option as 'false'. A turn is initiated only when the Player piece is switched to the player moving next, upon completion of the present move. When in a cul-de-sac, ZoG can not complete its current series of partial moves and passes its turn without switching said piece; the opponent can thus not move either. ZoG benefits from the resulting stalemate, so it seems to seek such situations and happily have a go at one, if any is found. I gathered that much only after I had already added some macros and board definitions to improve things. One hopes it doesn't hurt, but you could always comment the relevant parts of the ZRF if they seem to impede the engine's AI.
| |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 16 Registered: 5-2016 |
Rats. That solved zilch. I can't go on any longer. Good riddance, too. Sorry for wasting your time, Sean. | |||
| Sean Duggan (Dream)
New member Username: Dream Post Number: 121 Registered: 9-2000 |
I'm sorry I haven't been more help. I've been away from the system for a long time... | |||
| Mohamed El Mokhtar Messaoudi (Noggluggoid)
New member Username: Noggluggoid Post Number: 17 Registered: 5-2016 |
Thanks all the same, Sean. As a last resort, I further cluttered the ZRF with a safeguard I previously deemed perfectly redundant. As far as I could tell, it seems to do the trick but I am still wary.
|
