In Patch 14.8 (last patch), we saw WAY too many 3* 4 costs, which is why we made a bunch of system changes. But how do we know that for sure, and how did we do?

One of the many things our insights team does is help give us clear metrics on things around the game, so we can understand it better and give ourselves a nice benchmark or framework to try to achieve. During Set 9 we were having a similar issue, so our insights team created a great dashboard that shows around how many 3* 4 costs we were seeing per game at all elos and high elo only. We used this to inform where we wanted to be.

Generally high elo will hit their 3* 4 costs more often. So for them we aim for approximately 0.2 per game on average. Low elo ends up hitting them a bit less often, so something like 0.18 per game is right on target.

At the time, we used Set 7 and 8 to inform these numbers, as they were generally very close. Set 8.5, 9, and 9.5 we started seeing a rise in these, getting all the way up to 0.42 per game in 9.5 high elo. This is one of many reasons we made some system changes for Set 10, which we managed to get back to our 0.2 target. Low elo was as low as 0.14 during Set 10. Set 11 started similarly, which made sense since the 4 costs weren’t really meta. Then the patch hit…

And yeah, you can see on the graph below, we shot WAYYY beyond what is acceptable. And it just rising as players kept optimizing around it. High elo reached all the way up to 0.65 per game, which is more than TRIPLE our intended amount. Even low elo reached 0.45 per game. So we had to take action, and it was going to need to be something bigger to hit it.

Our goal was to get it at least in half as I mentioned last week, so 0.32 or less for the first swing. The good news is, we accomplished that! Right now it’s sitting at 0.28, which is still above our goal, but MUCH closer to reasonable. From here, we think the current power of 4 costs is mostly likely the cause, and helping them get more in line while buffing some lower costs can help this as well.

Now I’m sure a bunch of people will go “Duh, stronger units get 3 starred more”. And you aren’t wrong. When 3 costs are meta that number can get as high as 3.0 or more for them! But it’s important that we set a clear goal for where the game should be, so we can make adjustments to keep the game there. That is just one of many things insights can help us do!

Tomorrow I’ll be back talking about ranked. I played a LOT of games this weekend for the first time in a while. Some ups and downs for sure. It was fun though, and that’s what matters. Until then, have a great week and take it easy 🙂