| |
BGonline.org Forums
An Open Letter to SHG Players
Posted By: W Womack
Date: Monday, 19 October 2009, at 8:55 a.m.
I am posting this here as a way to go on record with the larger backgammon community in an attempt to set the record straight and perhaps to solicit some additional aid from those that might have an interest in what is going on at SHG. It is prompted by a conversation I had with a room host and several players last night.
A few weeks ago I posted in this forum my analysis of SHG dice, showing that the number of expected doubles was well below what would be expected from random dice. That analysis was prompted by a posting by Neil Robins that referenced his concern over the dice at SHG. My original post meet with some skepticism, although I think there was enough data there that many readers felt it warranted further investigation. Mike Petch posted a follow-up message that confirmed that the frequency of doubles was well below the expect values.
In response to that SHG rolled out a new version of the BG program in some rooms, the version number of this was .28. Mike Petch and I collected a large amount of data by kibitzing tables at SHG. Mike analyzed the data and the actual values were very close to the expected values for the number of doubles, runs, and bearing in checkers from the bar. SHG has since rolled out two other versions of the BG program, .32 and .33. Mike is still working on the analysis of version .33.
That is the background, now to the point of this posting.
Last night I was accused of ruining the dice on SHG, producing biased data, being unqualified to analyze and draw conclusions about the data. Also I have been told by several people that claim to play over the board backgammon that they never see 4 doubles in a row and one person that claims to have played competitively for 31 years that he has never seen 6 doubles in a row. I would like to address these issues, and perhaps solicit input and assistance from those on this forum that I know have many years of over the board experience.
1) Since neither Mike nor I have any formal relationship with SHG or the ability to change anything there we cannot “ruin the dice.” SHG responded to data that indicate that their dice were not what was expected from a random data distribution.
2) I resent the accusation that I was biased in handling the data. When I did the initial analysis of the data I actually expected to find that the frequency of doubles, which was all I was looking at the time, would come out to be close to the expected value. I did nothing to tamper with the data, or to attempt to bias the results in any way. I firmly believe that this is true for Mike as well.
3) Regarding my qualifications, while I have some statistical training I am not a statistician (and I do not believe Mike is either). However, one does not have to have an exhaustive knowledge or even an expert knowledge of a subject to know when some things are incorrect. Were were trying to decide if a 600K rolls sample with doubles occurring at say 16.1% instead of 16.67% was possible from random data then I would agree that we needed to apply some statistical methods beyond my expertise to reach a conclusion. But when you are off my almost half you don’t need a sophisticated statistical test to tell you something is wrong. If I might be permitted an analogy, as a new student of the game I would never presume to tell the WC players on this site that their play or cube in almost any situation was wrong. On the other had when I see my opponent playing their opening 52 as 13/6 or stacking 5 or 6 checkers on their ace point early in the game to prevent leaving a blot I am competent enough to say those are not good plays.
However perhaps someone that does have the statistical or mathematical qualifications and is interested in this would be willing to take a look at the data.
To the good English doctor that questions my qualifications and claims our data is “crap” but “has better things to do with his time than look at the data” you sir are simply resorting to a type of ad hominem argument. It might be effective in a courtroom, but I think we should be above such arguments and deal with the facts.
4) To the room host who thinks that people should “vote” on the dice because SHG is a “social” and not a “professional” site I say that is fine. But if you do that don’t call the game backgammon, and make it clear to everyone playing there that the dice on the site are not truly random dice. I would suggest the name Hackgammon, but that is a little to close to Nackgammon. Just don’t expect the top level players that have started to play on SHG to stick around, or the site to be acceptable to the more serious online competitions.
5) And to all of those players with all of those years of over the board experience that have never seen “real” dice come up like this I will freely admit that I have virtually no over the board experience. But I would bet that most of the players that read bgonline and do have real, competitive, backgammon experience would tell you that yes you do get 3 or 4 and sometimes more doubles in a row at times with real dice. I bet some of them even have recorded matches where it has happened.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.