Luck-PR statistics project crowdsourcing: Play long matches to help me with data collection!
A
Anton Hodgetts 🇬🇧 (scaredycat)
Hello everyone,
I have been collecting data on matches played on Backgammon Hub for a statistics project. I currently have 273 matches worth of data, and have run statistical tests on the data revealing that contrary to how some players will complain, the player in a match with worse PR tends to have good Luck on average! This has been determined to a high level of confidence with a 1% significance level hypothesis test of mean total Luck. I have been writing up my results this week and realised I need just a bit more data for long matches (9, 11, 13) in order to accurately test whether this surprising result is more or less extreme as match length increases and generally to better visualise how much the skill and luck factors change with match length. I need more data to have more confidence in the results for longer matches, then I will publish the report.
This is how you can help with gaining further insights: I need 10 more matches for each of the match lengths 9, 11, 13. So all I need you to do is play a new match for of one of these lengths (please do NOT share a match you have already played more than a week before reading this post), and then share the match link on this forum! When you play a match with the intention of providing data for the project, you must share the match link regardless of the outcome otherwise this could skew the results from selection bias, there is no such thing as a match that "didn't count".
This is the quickest way for me to get the data while ensuring there is no repetition of matches in the data I have already collected. Combining this extra data with what I already collected, I will have at least 20 matches for each match length and over 300 matches in total, which will finally be enough to say something interesting about it statistically.
Anton
D
Denis-George Constantin (playhunter)
We know for sure that rolls here are randomly generated and that are verifiable. So in theory, the better you play (in comparison to your opponent), the luckier you will appear, because by playing better you give yourself more better rolls than he can give for himself (and at the same time you leave your opponent with more bad rolls than he can leave for you) ! Unfortunately a couple dozens matches won't be able to provide you anything full proof I believe.
Also, it matters the skill gap difference between the two players - if the skill difference is 10 points PR the luck swings will be much more infrequent than it would be if the skill difference is only 2 points PR. - Keep track of this too !
I don't know how you gathered your data, but I want to let you know (in case you don't know) that you can access any player profile by clicking their nickname, and once you are on their profile, you can view their played matches. In case you did not look at my profile already, please look at it because I have a lot of 13 point matches (also a couple to 11 points).
In fact, check all players profiles that played in this tournament, and you will gather a lot of 13 point matches data from their profiles as we played a lot !
https://backgammonhub.com/tournament/bgahub-1?standings
To access a player's profile from that link, click on Matches tab, then click on each match in part to view it, then when the match table opens, click on each player's name and Hub will get you to their profile page. Then click on 11 / 13 in "Filter matches" to only see results for matches to 11 / 13 points, then click on "View" for those matches that have a PR displayed in PR section (because the other ones which do not display a PR figure there means were not played on Hub but on another platform and you will not be able to download them). And that is how you gather a lot more matches for your experiment !
A
Anton Hodgetts 🇬🇧 (scaredycat)
I would be happy to be your opponent if you are struggling to find players who want to play a long match. Send me a challenge on weekends as I am busy during the week with uni.