|How to Improve|
I wrote this article in its original form for Kit Woolsey’s site, then called Gammonline, now called GammonU, in 2000. JellyFish 3.0 and an earlier version of Snowie were the bots available to backgammon’s general public then, so some of what you read here relates to them, though I have updated the article so that it includes XG and no longer assumes you are using JellyFish. The ideas here about how to study apply equally well to XG or any other bot assistants you may employ.
How to Study Your Own Matches
One day I was minding my own business kicking a couple of FIBSbots around, and a human kibitzer asked me why my experience points on the server hadn’t been going up much lately. I told him I had taken some time off to study the matches I had recorded electronically from FIBS. (I also told him he was very observant!) He didn’t seem convinced that it took that much time to run a few matches by JellyFish, then rush back to playing again before the dreaded withdrawal symptoms began in earnest.
Analyzing your own matches can indeed be a quick process if you just click through them quickly, perhaps looking only at your own errors if you have XG or Snowie, and let your bot do all your thinking for you. But why settle for that, when you are so much smarter? Here is the method I recommend for letting your bot assist you in your ceaseless quest for knowledge, enlightenment, and the ability to squash your opponents like so many little ants:
- Record the match as you play it (electronically is fine). Feed it through JellyFish, Snowie, XG or any other bot you may have that lets you run matches through it conveniently.
- Help your bot find your errors. Set its gotcha points (what you will have it define as errors and blunders) at no more than .03 and .100, since on its first pass it may not otherwise catch as much. If it flags something, run it up to XGR++ (3ply on Snowie) to see if it “reverses on appeal”. Also, check out any other choices that were close in the original analysis, since there may have been another choice it will end up liking better than either play!
If it ends up preferring your play on its highest level of analysis by a good margin, you can usually assume your play was OK and continue. If it’s close or you still have doubts, you might want to run a quick rollout to see if it confirms the decision.
If you think any or your opponent’s errors were of interest, run them up to a higher level of analysis also.
- Make the bot tell you about the errors it didn’t catch. As you go through the match, even if the bot doesn’t flag a play, that doesn’t mean you can’t learn anything from it. This is an advantage of running a match shortly after you played it, because you may remember difficult decisions that don’t show up as errors. Another way to ensure you look at those plays in your online matches is to quickly write down something like “Game 4, Move 12” on a scratch pad while you continue to play. Then you’ll know where to look for a play of interest, even if the bot doesn’t mention it.
What if you made a correct call that was right by over .100 by means of a mental coinflip? If so, you were just lucky, and you didn’t understand that position. Give it just as much attention as you would if it had been an error or blunder!
- Question the evaluation if you know from experience that your particular bot has issues in the type of game in question. Complicated backgames and containment endings are probably still the worst offenders, though XG does better with them than the earlier bots did.
- As best you can, distinguish among careless errors, matters of style, minor nuance plays, and true concept errors. This last group is the one you need to worry most about! Concept errors tend to involve game plan selection, a fork in the road where you make a choice from which there will be no turning back. If the error is large (I consider > .70 equity large), I sometimes find it helpful to print it out, then review it 3-6 months later along with others I have been collecting.
- Run evaluations for the same play at double match point (assuming the play occurred at some other score in the actual match) and also as a money problem. These cases help you to see how gammons, cube access by one side or the other or both, or the lack thereof, affect the decision in question. (If you have never done this, you can save a position from an XG match using ctrl-shift-S.) You can also run other scores while you’re at it, of course.
- If you feel it is appropriate and convenient, do a rollout on strong settings of at least the original position to ensure that the bot understood it well. You definitely need to roll out long-horizon endgame problems, and any problem where the evaluated difference was relatively small, but it doesn’t appear to you to be a “matter of style”.
It’s not usually necessary, but you might want also to “spot check” a couple of key plays immediately following the original one, to be sure the bot is not making massive errors in the rollouts. This can affect the results big time, especially if one side is much more difficult to play than the other. Fortunately, with XG this is much less often a problem than it was with the earlier bots.
- If there is still a significant difference between the best play and the second choice after the rollouts, determine what features of the position are most important. Do this by changing one feature at a time and seeing what effect it has on the play selections.
- Propose a general principle governing the type of play you have just studied. If possible, come up with one or more radically different positions where you think the decision should be similar, based on your new principle, because the key features are similar to those of the position you studied. Ask the bot if it agrees.
Isn’t this all kind of time-consuming? Well, uh, yeah, it is. If he’s reading this, now that FIBS player who thought my experience points weren’t going up very fast knows why. But for me, part of the “experience” gained from playing is learning as much as possible from my games. Backgammon is so much more satisfying, at least to me, when instead of settling for the intermittent reinforcement that the dice often provide, I also have a sense of real progress.
About the Author
Mary Hickey learned to play backgammon at the very end of the 1970s craze, then dropped out to pursue other interests and raise a family. She returned to the game at the end of 1995, and has been playing, teaching, and writing about backgammon ever since.
Mary was the U.S. Open Champion in 2010 and repeated the feat in 2011. She finished sixth in the 2010 American Backgammon Tour and ninth in 2009. Besides her back-to-back U.S. Open victories, her tournament wins include the Mid-Atlantic Championship in 2010; the Ohio Open Championship in 2009; the Ohio Masters event in 2002, 2003, and 2006; the Indianapolis 300 masters event in 2006; and the Blitz in Las Vegas in 2006.
Mary Hickey and Marty Storer are the co-authors of What’s Your Game Plan? Backgammon Strategy in the Middle Game, published in 2011. Mary began working with game-plan selection problems many years ago, and published several of them in 2006 on the GammonLife.com website. As of this writing, this site is still active and includes Mary’s columns, along with other backgammon news and information, tournament reporting, player profiles, and photos.
Mary also writes about backgammon for the online magazine GammonVillage, where she continues to report on live tournaments and share the occasional tutorial. Her series of columns was released as a book on CD in 2010, and the columns plus her other work for them continues to be archived at the GammonVillage.com website. She also writes a monthly column about the early game, titled “Starting Off Right,” for PrimeTime Backgammon, the magazine of the U.S. Backgammon Federation.
She has also published numerous articles in the Chicago Point, including several for the “Your Move” column. Her work for the Flint Area Backgammon News includes quizzes, articles, and book reviews. She was also a contributor to the recently discontinued Hoosier Backgammon Club Newsletter.
Mary’s other current writing projects include some speculative and experimental fiction, a book about the experience of moving from the city to the country, and a book about time management. She is the author of a book about Mary, the mother of Jesus, titled Arise and Call Her Blessed, published in 1999 by Paulist Press.
Mary Hickey plays online at FIBS, Gridgammon, and Safe Harbor Games as “mamabear.” She also teaches the game live and online. She lives on a “mini-farm” in central Ohio with her husband Bob.