2017 By The Numbers

Posted by Antibody on January 21, 2018

We said goodbye to the year 2017 quite a while ago, so now it is time once again to see what that year’s worth of data had to tell us. As I did last year, I unleashed Pandas and Matplotlib on 2017 data from the database that powers XonStat: XonStatDB. Let’s see what it has to tell us!

Let’s first look at the number of games throughout the year:

The number of games for the first part of the year - through July - is lower than the year before, in most cases only by a little. What’s interesting is that the rest of the year has higher numbers! Unfortunately they weren’t enough to make up the huge two months we had in January and February last year, so the overall amount of games through the year is down by just over 19K. The discrepancy might be explained by the heightened interest in the game through all of the practice sessions, YouTube videos, and player interviews for the Xonotic World Cup that was going on at the time. Perhaps it is time to do another one of those!

Next let’s check the most popular time to find games:

Once again 8PM UTC takes the cake for the most popular time to play. This corresponds to the evening time for EU residents, so it appears that they enjoy de-stressing by fragging. As far as days go, Friday is unsurprisingly the best one. Everyone is getting their weekends started right!

The next chart shows the number of distinct players per month.

As with the number of games, the number of players is down for the first half of the year compared with 2016. However, unlike the number of games measure the overall number of distinct players for 2017 is much higher! Approximately 3K more players played in 2017. It follows that the bigger the pool of players that we have available to play in games, the more games we’ll end up having, so let’s do our best to welcome these new players into our community so we can have an awesome 2018.

Next let’s have a look at weapons:

We can’t extrapolate a ton of information from this chart given that it contains data from all game modes and all configurations, but it is useful to get a high level view of what weapons players are using. On that front it is good to see that each weapon has a place or “band” on the vertical bars, which indicates that all of them are being used. We’d have a concern if any of the “core” weapons were only represented with tiny slivers.

As a fun side note to this chart, please be aware that the blaster metrics are excluded. When I ran the numbers this time I saw some truly whacky numbers for that weapon that I’ve yet to explain. I’ll defer that explanation to another time so I can do some analysis! It’s likely that some server was misconfigured. Who knows - someone could have hosted a week long blaster-only tournament. The jury is still out on this one.

Moving on to the frags dealt by all of those weapons:

This one also has representation from each of the weapons in the game, which is good to see! One point to note is that the devastator almost always leads in frags versus, say, the vortex. A common complaint is that the vortex is overpowered (op), but the data both on damage and frags indicates otherwise, at least in the general sense. More digging into different modes may yield different conclusions.

That’s it for this year’s data exploration. As mentioned before, I’ve placed the code that generated these charts on Github if you’re the coding type. If you’re not, I hope to see you out on the servers instead - let’s get out there and do our best to bump up the 2018 numbers!

Previous post: phrantic walkthrough

Next post: XonStat Records its Millionth Game