As much as game design is about art, it’s also about numbers. UETF contains some very specific metrics that I wanted to track as indicators of choices players were making. Here are the results for release one, how I’m interpreting them and if/how I’m using the data to improve your playing experience.
Players Levels Gained:
There is some very important data in this graph. Note how it’s heavily skewed toward the left side – this is expected since a player must gain the lower levels to get to the higher levels. Still the dropoff is pretty dramatic: less than 25% of players get from level 1 to level 4. With a max level cap of 10, this means that more than half of the players are only unlocking 20% of our class content, meaning most players abandon their characters after only 2 levels. As a goal, I’d like to see over 50% of players see over 50% of the content. This has lead to a couple changes in the 1.1 build: level requirements have been significantly reduced, and the level cap has been reduced to 9. Hopefully making leveling easier and increasing the upgrades per level will encourage players to play for a longer amount of time.
Percent of Games Completed:
34% of all games started completed (success or failure)
Again, an extremely important statistic. Basically, only 1/3 of games are played to completion; 2/3s are abandoned; a serious failure to engage the audience. Due to this, for the 1.1 update I’ve reduced the max wave level to 9 and also changed the game-play to reduce the amount of ‘boss’ monsters and add more ‘mooks’ (mooks being weak creatures designed to be cannon fodder). I think it will help to bring up the pace of the game and also increase the enjoyment players feel from getting some easier kills.
Character class selection by percentage:
Rangers: 53.3%, Fades: 25%, Grunts: 21.66%
Tracking this was mostly out of curiosity, but there is an interesting trend here; overwhelmingly, people are choosing rangers – more than the other classes combined. My guess is that it’s because it’s the class displayed by default; randomizing which class is selected in the display menu may remove this bias, or people may just love the Ranger class, something we’ll be experimenting with in the 1.1 release.
Percentage of Gained Character levels by Class:
Ranger: 40.6%, Fade: 33.8%,Grunt: 25.4%
So the skewing in this stat is a reflection of the last stat in the Ranger bias. Also, Grunts are lagging a bit behind Fades – not by a significant margin, but something I’ll keep an eye on.
I hope the stats provided in this section have proved interesting, at least – check back at the next release to see how our changes fared!
Subscribe
