Is Going to Tournaments Correlated with Winning Them?

I was grabbing a current snapshot of the current ITC standings (as of Jan. 29, 2017) for another project, but while I was looking at the data I started wondering how correlated tournament attendance is with ITC score.

So I decided to find out.

Now, clearly there’s some relationship that’s inherent. Your ITC score is calculated as follows, from the Frontline Gaming website:A player’s ITC Championship Rankings Score is comprised of their top 5 scores from the following categories:

• A player’s ITC Championship Rankings Score is comprised of their top 5 scores from the following categories:
  • Up to 3 Rogue Trader Tournaments
  • Up to 4 Grand Tournaments
  • Up to 5 Major Tournaments

That means with less than five tournaments, there’s an effective cap on your score.

Also, as a disclaimer, there’s some flaws in the data. I know this because I looked up myself and I showed up as two different entries. Undoubtedly, I am not the only person this is true for, and I haven’t bothered to figure out why that is, or clean and merge these split players. But I also don’t think this will massively bias the results.

With those caveats aside, lets consider the data.

Events Attended: The vast majority of people entered in the ITC system only went to one event. Again, this is overstated a little (my two events show up as two one-event players) but the median number of events is one, and the mean is 1.90. The maximum? 24. Good on Joshua Death for that.

Correlation Between Score and Events:

First, lets look at the picture:

There’s a clear relationship for those first five events, obviously – each one of them can only possibly improve your ITC score. After that, the spread widens a fair amount – it looks like there’s a little bit of a relationship, but you can get some seriously high scores with less than 10 events.

Quantifying this relationship, I fit a quadratic curve to the data (basically, predicting Points based on Events and Events^2):

Numerically, there’s a very strong relationship right up to about 10 games, wherein there seems to be some diminishing returns in going to more games, or at least no marginal benefit of doing so.

Out of curiosity, I also looked at this for a trimmed data set, where everyone with less than five events were removed. This group can be thought of as the people in the ITC tournament who could actively have been considered to be “going for it”, as anyone taking the ITC seriously basically has to go to five events of some sort, and this cuts out people like me, who went to a couple ITC events essentially by accident to meet people and play some games. This does cut the dataset from more than 5000 entries to less than 500.

There’s still a very significant relationship in the data. Again, pictures:

The confidence interval gets way wider, but again – you can see there’s a pretty steady gain until about 12 to 15 events, at which point it fades away.

What can we take away from this? Well, I think there’s a couple things:

1. There is a luck element to tournament games – a bad matchup, getting seized on at a bad time, or just a game that didn’t break the right way for you. Going from five games to six or seven can help smooth out some of that luck.
2. Practice does indeed help.
3. While practice does help, as do more “swings”, there comes a point where just playing more games doesn’t necessarily manifest as higher scores. This makes sense – I don’t think anyone would argue that the higher echelons of tournament play is entirely a random selection of players, so there’s no reason to believe that just more games means a higher score.

There’s also a confounding result here – serious tournament players (people who presumably might win tournaments) presumably also go to more events. Getting rid of the casual sub-5 event players might even this out somewhat.

So if you’re looking to maximize your ITC score for next season? Maybe add in another tournament or two to your travel plans, but don’t go overboard and try to flood the series with results – there’s definitely a cap where more games aren’t going to effectively boost your score.

This does however suggest that in my next round of predictive modeling for tournaments I should be including the number of events someone has intended…stay tuned.

Enjoy what you read? Enjoyed that it was ad free? Both of those things are courtesy of our generous Patreon supporters. If you’d like more quantitatively driven thoughts on 40K and miniatures wargaming, and a hand in deciding what we cover, please consider joining them.