Imperfect Symmetry: A Deep Dive Into Control / KotH Maps in OWL
Examining patterns across the different control maps and teams reveals some interesting results. For the sake of simplicity, none of the below analyses consider team composition and combine results across stages. I also use KotH and Control interchangeably to describe the same thing specifically to trigger anyone who decides to read my work.
Tl;Dr — There is statistical evidence that would suggest that some rounds of some control maps have a higher chance of appearing first than others, overall capturing the point first nets you a fifteen percentage point increase in chance of winning, different maps have different value associated with capturing the point first, and some teams are better than others at capitalizing on the first capture.
Unlike all other game types control creates a situation in which both teams have equal resources at the start of the round. The sides are symmetrical, there is no spawn advantage, and the objective is stationary. There are only two possible advantages a team can have on a Control map — Control of the point and Ultimate Advantage. Ultimate Advantage is a pretty ubiquitous concept in Overwatch, so I will instead choose to focus on point control.
Without doing a deep dive, any one of you could probably guess that the team that is in control of the point has an advantage, but how do you quantify that? Does the map matter, does the team matter? These are questions I aim to answer below, and the results I think are fairly interesting.
Let’s start with a simple first inquiry — is the order in which the Control rounds appear really random? Each Control map has three separate rounds, for example Nepal has Village, Shrine, and Sanctum. You would expect that each of these three rounds would have a 33% chance of appearing first, but that is not necessarily the case. Time to lay out some numbers.
Ilios has been played 74 times in the OWL, and each round has appeared first between 29% and 36% of the time. The range is similar for Oasis (played 74 times) which has a range of 31% to 35% — both of which fall within an expected deviation away from 33%. However, when you look at Lijiang Tower and Nepal, that parity goes away. Shrine has appeared as the first round 40.2% of the 67 times the map has been played in OWL, but this deviation (from an expected value of 33%) is not significant. That changes on Lijiang Tower though, where Garden has a 48% of appearing first which is a statistically significant difference (P = 0.012) compared to the expected population proportion of 33%. I want to be careful here about what the takeaway is from this finding. Statistical significance implies an arbitrary degree of certainty that an observed value is different than an expected value (in the one-sample test), and does not prove beyond any doubt that there are non-random factors in which round appears first on a control map.
Addressing the impact of this inequality of first rounds is a bit tough. Firstly, its fair to point out that what really matters is whether a map-round appears in the first two rounds or not as those rounds are guaranteed to be played no matter what. And if you use those terms to examine the discrepancy, the statistical significance goes away. But Overwatch is a game of momentum, losing round one could impact your performance in round two. The last thing I’ll say is this — the Gladiators win on Lijiang Garden 25% more than their average Control round and the Shock lose 23% more than their average on that map, so take away from that what you will.
Moving on to the heart of the article, what does capturing the first point actually give a team in terms of an advantage. Across all maps and teams, the team that captures the point first goes on to win the round 65% of the time. However, as you can see below this percentage varies fairly heavily depending on which map/section you are playing.
Many of these deviations make perfect sense. Control Center on Lijiang has the highest win rate increase from capturing the point first, and that map has some of the best holding positions. Think about how often you will see teams push up to the outside entrance of the other team after capturing the point, and how oppressive that hold is. Likewise Oasis Gardens has the lowest impact of first point capture, barely increasing the odds at all at 53%. Although Oasis Gardens has some nice high ground, there are many routes of access to the point and defending it is often tough to do due to enemy flankers.
Of course, there is also a big difference in the way different teams play and perform on control points. It has often been said that KotH is the map type where mechanical skill is the most important compared to strategy — and there is probably nothing more pure than the first fight on a control round — both teams come out of the gate at the same time and have the same distance to cover and have the same ultimate charge. The Fusion has long been lauded as a team with high mechanical skill — I mean we beat the Outlaws in our first game in Stage 1 when half our team had literally just gotten off of a plane from another country — and as you can see below, that mechanical skill shines in terms of the first fight win rate (determined by who captured the point first). The Fusion are the best team in the league at capturing the point first on KotH, and the Spitfire actually capture point first less than half of the time despite being a top team in the league.
The red line in the above chart tells a different story. Where the red line is above the dotted black line indicates a team that is better than average at converting the first point capture into a round win. This is not to be mistaken for ‘Win Rate Given First Capture’. For example, the Shanghai Dragons have the highest ‘Winrate Boost’, but that is partially due to the fact that they have an overall winrate of 16% on Control rounds, and therefore their boosted rate of 40.7% seems quite high in comparison. The chart below shows how teams generally fair on KotH rounds. Keep in mind this win rate is not the same as KotH maps won.
For no reason other than curiosity, I wanted to look at which teams had the most ‘shut outs’ or ‘perfect rounds’ in which the enemy team did not score a single capture point. The charts below show that in fact the New York Excelsior do not have the most shut outs, instead it is the Boston Uprising sitting pretty at #1 with 20 perfect rounds. Equally impressive though is that the Uprising are also the team that gave Shanghai Dragons their only perfect KotH round on Ilios Lighthouse. The chart on the right displays which KotH rounds are the most subject to ‘perfect games’. Interestingly, Oasis Gardens is the biggest blowout round even though it is also the round with the smallest impact on win rate from capturing the point first. Every day really is an adventure!
Conclusion & Discussion: There are no real actionable takeaways from an analysis of this level of depth, but I do think its interesting to examine. I will be the first to admit that ‘capturing the point first’ does not 100% correspond to winning the first fight. There are many examples where a team will manage to cap the point as it is getting cleaned up and only come away with a measly few percentage points of progress. Additionally, it was hard to find the right way to measure a teams relative success at converting first fight wins into round wins. I wanted to separate that effect out from their baseline KotH win rate, but I am sure there are better methods to accomplish that than the one I used.
What do you all think? Did I miss something big? Are you surprised by the results?
