Preseason 5 patch analysis – How has 4.20 changed dragon’s effect on games.
With the new season upon us and major changes in the pre-season patches I wanted to analyze how these are effecting the game. This is tough however given there are only limited games on this patch, so we narrowed our scope this time just to look at how the reworked dragon has changed.
The first thing we looked at was how the first dragon win rates changed. To do this we compared all of 2014 LCS data (both NA and EU) to the 4.20 data we had which was OGN matches, IEM San Jose, and NA LCS Expansion.
Some key things to note:
- First dragon over 2014 season has been getting more and more important. While first tower and first blood were rather flat in trendline over the 2014 season, first dragon had a clear upward trend, making it the most important early “objective” nearing the end of the season.
- There is a wide variance from patch to patch. There is no one definitive number when it comes to analyzing these games and even with decent samples (anywhere from 60 to 150 games on each patch).
- First dragon had a slightly lower win rate in the 4.20 patch than it did at Worlds in the 4.14 patch, and significantly lower than the trend line for first dragon.
So, while this is only a moderate sample it has shown that this new dragon change hasn’t completely changed the importance of early dragons, it continues to be in the same rough range of win rate as it was in previous seasons. Instead the big upward trend was First Blood after the early game kill & assist bonuses were reverted back to full gold value in the first few minutes.
Power of consistent dragons
Another aspect of dragon to look at was now that dragon has a stacking buff how did consecutive dragons work in the previous season compared to now. We had a large sample for the 2014 data, it includes all of NA and EU LCS (including playoffs) leaving a lot of data to get an interesting curve of the power of dragons. This data, along with a look at the win rates of individual dragons: http://puu.sh/dFFCQ/630ab7a31b.png shows in my opinion that the 2nd dragon was the most important. It is the biggest jump from the first to second win rates (62% ->77%) as well as having a 72% win rate on its own. This made 2nd dragon a more telling winner than the first, which might have been traded or cheesed in a 2v1 situation. However, in the new patch this has changed more so as the getting the first and second dragon sequentially gives almost no increase to win rate, and the win rate of the second dragon is down to 57%, even lower than the first dragon.
These were the big changes to me in this patch, that the first dragon is equally good but the second is much less important. Or it signals some kind of weird rubberbanding where the second dragon is often got by a different person than the first.
So how long does it take to adapt to the new meta?
While these conclusions outlined above are tentative due to the small nature of the sample size, it also will be a while until we see consistent results. A major change last season was the vision rework, with trinkets and pink wards changing significantly. Below is a look at how many pink wards on average were bought per team in each patch (for NA & EU LCS).
This I think shows that it took a bit of a learning curve before the meta really became set and the changes were fully fleshed out in the meta game. For these pink wards it started hitting its trend at ~14 pink wards came at around 4.4, which was the last week of Spring LCS, almost 4 months after the pre season change. As such even as we start to get real data with the start of LCS it might take a few months before we fully realize the changes that have occurred to this preseason, especially in regards to the tower and dragon changes.
Data collection: Spellsy
Graphs with help by: Benson Li
Help associating weeks to patches by esportspedia.