EPT1. Introduction
I think not enough people are discussing about the difference between skills and performance. What is a Bonjwa? The most skilled player or the player with the best performance (i.e. results)? And if we talk about results, we need to assign a weight to online vs offline performance, how much weight does each carry?
Measuring skills is hard because they it is subjective. Performance (results), on the other hand, are much more quantifiable. Here I compare Soulkey's performance against the top players of each race, with the caveat that performance is an imperfect indicator of skill. For instance, Shuttle is an ASL champion, while Snow is not. In the recent SSL, Action got eliminated by Sea, and in ASL17 Light got eliminated by Shuttle, and Queen, an ASL double champ, didn't even qualify.
Note: I won't try to convince anyone about how to define who is a Bonjwa, nor of the relevance of online vs offline performance. That would be a fruitless uphill battle.
2. Stats and graphs
2.1. Elo score
In a previous post I showed that by far the best performing players up to 2023 had been Snow, Light and Soma, with Soulkey joining them starting in 2024. For completeness sake, I share an updated version of that figure, comparing the 28-day moving-average Elo score of the top two players of each race (if you are not convinced they are the best players of each race, check the post linked above)
Top two/three players of each race:
+ Show Spoiler +
![[image loading]](https://i.imgur.com/D1X0vpo.jpeg)
Now let's remove the clutter and focus on the top dogs: Snow, Light, Soma, and Soulkey:
![[image loading]](https://i.imgur.com/W0ppOHK.jpeg)
As we can see, Soma used to perform much better than Soulkey, until he started his military service in December 2023 (?). My takeaway from this is that Soulkey still needs to prove himself as the top Zerg once Soma returns.
2.2. Win rate per match up
However, the Elo score doesn't give us a full picture of the performance of a player. So I calculated the win rate % against each race in moving windows of 4 months (where the curves give big steps, the sample size is very small, so the values in those regions should be taken with a grain of salt.
To make the figures below (matchup win %), I only considered the top 16 players, trying to keep a balanced proportion (5P, 5T, 6Z):
- Protoss: Snow, Bisu, Mini, Best, Stork.
- Terran: Light, Rush, Royal, Sharp and JyJ.
- Zerg: Soulkey, Soma, Queen, Hero, Action, Jaedong
- Light:
![[image loading]](https://i.imgur.com/umzuCSr.jpeg)
- Snow
![[image loading]](https://i.imgur.com/TrG7u59.jpeg)
- Soma
![[image loading]](https://i.imgur.com/jcaL4Ll.jpeg)
- Soulkey
![[image loading]](https://i.imgur.com/NoBHsX4.jpeg)
The first, and expected, takeaway is that Light and Snow have shown a level of domination in the mirror matchups that Soma and Soulkey can only dream of. Note: Soma's high win rate % against Zerg in the beginning of 2023 is due to the small sample size, e.g. if you win 3 out out 4 matches, your win rate will be 75%, however this curve is later stabilized in the second quarter of 2023 (the number of ZvZ is very low).
The second takeaway is that Snow consistently dominated all matchups during 2023. In 2024, he has continued to dominate in PvT and PvP, however, his win rate against Zerg has just been slightly above average, if not average. Light and Soma on the other hand, only managed to dominate in two of the three match-ups in 2023. In 2024, the TvT data is too sparse to arrive to any conclusions. Light has improved a lot his TvP during 2024, however, his TvZ dropped dramatically during the first half of 2024, only to recently rise again.
The third takeaway is that Soulkey's domination started only in the second quarter of 2024. From then on, he has improved tremendously in ZvP, with a never-ending upward trend. His ZvT winrate saw a huge increase in Q2-Q3, only to drop to barely around 55% in the last two or three months. People have been attributing this to the map pool change, however, I am not completely convinced. Let's take Radeon as a benchmark, one of the most balanced maps. Since August 1st (roughly when Soulkey starts to drop his performance against Terran), he has won only 1 out of 12 matches against Light in Radeon. I won't delve deeper on the impact of the map pool on the match-ups win% because I don't have time for that.
2.3. Win rate against other top players
Again, for completeness sake, I calculated the win rate against each of the top 16 players mentioned above (ignore the mirror match-ups because the sample size is too small and thus unreliable). For each player, I only considered opponents against whom they had played at leas 100 games. The data are somewhat messy and erratic because of the small sample sizes within each window, so I will just share the figures with a spoiler tag to avoid overwhelming the reader with messy figures unless they choose to see them.
- Light:
+ Show Spoiler +
![[image loading]](https://i.imgur.com/QW9fx43.jpeg)
- Snow:
+ Show Spoiler +
![[image loading]](https://i.imgur.com/hP5oD4T.jpeg)
- Soma:
+ Show Spoiler +
![[image loading]](https://i.imgur.com/YWoSBM9.jpeg)
- Soulkey:
+ Show Spoiler +
![[image loading]](https://i.imgur.com/fAokT8r.jpeg)
Notably, Soulkey has a <50% winning rate against Light in the recent months. Weirdly enough, no one is talking about it. Soulkey maintains a winning rate > 50% against all other players with a big enough sample size to be considered here. Interestingly, his win rate against Snow in the last 2 or three of months is high (55-62%), but not crazy high as we would be led to believe by the results of SSL.
![[image loading]](https://i.imgur.com/lYRZuP9.jpeg)
