Forum Archive : GNU Backgammon

Even-ply/odd-ply effect From:   Scott Steiner Address:   nospam@nospam.nospam Date:   13 December 2002 Subject:   Gnubg and other bots: odd-ply vs. even-ply Forum:   rec.games.backgammon Google:   3DF9FEE9.7D510825@nospam.nospam I have a question concerning gnubg's analysis' ply level. I suppose it is a general question that affects all bots, not just gnubg. Reading posts on this NG, I've heard several times that gnubg's analysis at 1-ply can be unfortunate at times and produce not so good results and that one should rather use 2ply beacuse it gives much more accurate results, which I do anyway. Further, I got the impression that odd-ply analysis is not so good as even-ply, so I ask if there is any truth to this and briefly how to explain this. For instance, I usually analyse using gnubg's preset worldclass++ setting i.e on 2-ply. On some moves where I have a dispute with gnubg's 2-ply analysis, I evaluate the top moves or the cube decision at 3-ply and see how my play ranks there. Sometimes I would go up to 4-ply, although very rarely because that takes too long. I notice quite often that the even-ply analysis are usually close and agree among each other but they disagree with the odd-ply analysis i.e 2 and 4 ply would agree but 3ply gives different results. Actually I noticed that quite often when going through some examples of Robertie's Advanced Backgammon. So, I guess the question is what is behind this odd-ply and even-ply analysis level and which should I trust more and what pitfalls are there when using either one. thx! Gary Wong  writes: You are quite right that several types of positions exist where odd- and even-ply evaluations disagree. Unfortunately, there is no general procedure to decide which one to trust more; in fact, if the two disagree by a significant amount, then that is a good indication that you should be suspicious of both. A long time ago, I found some results which made me suspect that 1-ply evaluations were _worse_ than 0-ply for an old version of gnubg. However, a subsequent longer run of tens of thousand of games indicated that 1-ply was indeed stronger by a small but significant amount (+0.012ppg cubeless, +/-0.004). On average, you can expect n+1 ply evaluations to be more accurate than n-ply evaluations; there is nothing to suggest that even or odd n is better overall. Cheers, Gary.   Did you find the information in this article useful?           Do you have any comments you'd like to add?

 

GNU Backgammon

Analyzing GamesGrid matches  (Roy Passfield, Dec 2001) 

Batch analysis tool  (Øystein Johansen, June 2004) 

Cache size  (Ned Cross+, Mar 2004) 

Compiling for Windows  (Øystein Johansen, Jan 2002) 

Edit mode removing checker from bar  (Scott Steiner+, May 2003) 

Entering an annotated match  (Albert Silver, Dec 2003) 

Error rates: Gnu vs. Snowie  (Raccoon, Mar 2006) 

Even-ply/odd-ply effect  (Raccoon, Nov 2004) 

Even-ply/odd-ply effect  (Tom Keith+, Oct 2003) 

Even-ply/odd-ply effect  (Scott Steiner+, Dec 2002) 

Filter settings  (Robert-Jan Veldhuizen, Nov 2004) 

Gnu 0.13 versus Jellyfish and Snowie  (Torsten Schoop, Aug 2003) 

Gnu 0.13 vs. Snowie 4  (Albert Silver, June 2003) 

Gnu 0.14 vs. Jellyfish  (Michael Howard+, July 2003) 

Gnu versus Snowie and Jellyfish  (Michael Depreli, Oct 2005) 

How luck factor is calculated  (Gregg Cattanach, Aug 2002) 

How rollouts work  (Gary Wong, July 1999) 

How to enter an illegal move  (Øystein Johansen, Aug 2003) 

Importing .gam files  (PAR+, Mar 2005) 

Importing PartyGammon matches  (rew+, July 2006) 

Improving your game using GnuBG  (D.U.G.+, Nov 2002) 

Installing on Windows  (maareyes, Oct 2001) 

Interpreting JSD's  (Adrian Wright+, Feb 2005) 

JSD's and confidence intervals  (Daniel Murphy+, Jan 2005) 

Logging rollouts  (Øystein Johansen, Oct 2004) 

Luck rate  (Kees van den Doel+, May 2002) 

MWC versus Equity (EMG)  (Ken+, Apr 2005) 

Manually entering first roll  (Andreas Graf+, Apr 2005) 

Match equity tables  (Raccoon, July 2005) 

Personal reflections  (Louis Nardy Pillards, Sept 2002) 

Playing two computers against each other  (Stanley E. Richards+, Mar 2008) 

Python scripting  (Øystein Johansen+, Nov 2004) 

Quasi-random dice in rollouts  (Ian Shaw, Mar 2004) 

Question marks in game list  (Jim Segrave, July 2005) 

Questions and answers  (Jim Segrave+, Jan 2003) 

Questions and answers  (Jørn Thyssen, Aug 2002) 

Restarting a rollout with different settings  (Jim Segrave, Apr 2005) 

Restarting a rollout with different settings  (Robert-Jan Veldhuizen, Apr 2004) 

Rollout settings  (geoff arnold+, Apr 2007) 

Rollout settings  (Stick+, Nov 2005) 

Rollout settings  (Robert-Jan Veldhuizen, Mar 2004) 

Rollout settings  (Ian Dunstan, Aug 2003) 

Rollout settings for the impatient  (Robert-Jan Veldhuizen, June 2004) 

Running rollouts in background  (Bruce+, Apr 2004) 

Saving rollout results from command-line interface  (Jeremy Bagai+, Apr 2006) 

Saving rollouts  (Mislav Radica+, May 2006) 

Setting GnuBG's playing strength  (JP White, Sept 2001) 

Setting skill level  (Jim Segrave, Apr 2004) 

Setting up and saving a rollout  (Albert Silver, Dec 2003) 

What's GNU?  (Gary Wong, Oct 2001) 

Which player is player 0?  (Neil Kazaross+, Oct 2004) 

From GammOnLine         Long message         Recommended reading         Recent addition

 

Book Suggestions Books Cheating Chouettes Computer Dice Cube Handling Cube Handling in Races Equipment Etiquette Extreme Gammon Fun and frustration GNU Backgammon History Jellyfish Learning Luck versus Skill Magazines & E-zines Match Archives Match Equities Match Play Match Play at 2-away/2-away Miscellaneous Opening Rolls Pip Counting Play Sites Probability and Statistics Programming Propositions Puzzles Ratings Rollouts Rules Rulings Snowie Software Source Code Strategy--Backgames Strategy--Bearing Off Strategy--Checker play Terminology Theory Tournaments Uncategorized Variations