‘Chiu on This’ is a short and regular opinion blast
There has been an interesting discussion going on about the idea of MVPs. This topic has become more relevant as we are in the Astralis era. The Astralis team is crafted in such a way that there is a seemingly perfect balance across all roles to the point where every player on that team has ludicrous stats. After EPL8, there was some discussion on social media as to whether or not dev1ce or gla1ve should have gotten the MVP award as gla1ve had a huge performance in the second map of the finals that helped close out the series at 3-1 instead of going to a potential fifth game.
Thorin did a video discussing this topic and how he models the MVP award in this video. This was followed by a reddit post by kazcmot talking about why he doesn’t agree with the ideas posited by Thorin from his video.
As I have my own views and ideas about the topic, I’ve decided to do a series of blogs explaining how I model the MVP award by using this particular tournament. I consider this particular issue is an incredibly exhaustive one as you have to fundamentally question every presupposition before you can break down what it means to be MVP.
Before we start though, I personally don’t think there is a truly objective way to do it that can be universally agreed upon. When I start to break down all of the criteria and models they are founded on, you’ll understand that there will be multiple points where regardless of what system someone uses, there will be multiple subjective inputs that have to be plugged into the system. For me, the only valid ways of approaching the MVP is to map out the criteria, list them, and then consistently follow them.
The last part is hard as there are going to be cases where it’s extremely close. To explain why, I’ll list out my own criteria. I think the criteria needed to judge an MVP are as follows. They need to have the highest level of consistency throughout the tournament. Secondly, their run needs to be contextualized within the context of what they do on their team. Third, their role needs to be contextualized relative to other MVP candidates. Fourth, each MVP’s path in the tournament needs to be mapped out and weighed against each other. This is critical in the case of MVPs on different teams as their teams will have faced off against different opponents. This could potentially create lop-sided opponents or create particular matchups that may be weighted higher than the other. Finally, each match on the way needs to be contextualized in such a weigh that it takes into account the relative value of a particular match.
The first is something that most people agree on and this is reflected by the fact that people put so much weight on HLTV stats as they are the quickest way of evaluating a player’s performance throughout a tournament.
As for the second piece of criteria I pointed out, to figure this out you have to understand the essential strategies of each team. Every team has a polarizing player on it, someone who is enabled to carry the game. Recognizing that fact allows us to roughly estimate the amount of responsibility that player has to carry on their shoulders. The most obvious example is s1mple who is doing around 70-80% of the heavy lifting of that team.
From there, we must then compare the MVP nominees to each other through their roles on the team and through the context of the players they have on their team. I explained earlier that s1mple is a polarizing player, but so is NiKo on FaZe. However if there came a tournament where the two were nominated for MVP status, s1mple would have an advantage in the direct head-to-head because he plays with Edward and Zeus and thus he is essentially carrying a heavier load. There are potential factors that can mitigate this though if both Edward and Zeus have a great tournament while someone like Olofmeister and Karrigan have a bad tournament on the FaZe side. Thus in order to ascertain a MVP candidate, on some level you must also evaluate the level of their teammates throughout the tournament as well.
The third piece of criteria is to map out and contextualize their opponents. For instance, in the PGL Krakow Major we know that both Immortals and Gambit made the finals. However each team had a wildly different path to get there. In the case of Gambit they beat: Mouz, G2, and VP to get out of the groups. Immortals beat: Vega Squadron, Na`Vi, and Flipsid3 in the groups to get out. In the playoffs, Gambit beat Fnatic and Astralis to get to the finals. Immortals beat BIG and Virtus.Pro.
It’s fairly clear then that Gambit had a far harder road to get to the finals relative to Immortals and thus if there was a MVP candidate on the Immortals side, that would weight against them in my final accounting as they played against easier competition relative to Gambit in that run. As for how I do it, I generally just use a tier system which is similar to the idea of the CS:GO kingdoms tiers.
The final piece of criteria is breaking down the value of each match. For instance, it’s far more impressive to do an incredible job in the finals of a tournament than it is to be good in the group stages. On the other end, it’s far more impressive to be consistently great throughout the tournament while be average in the finals compared to being average through most of the tournament and being great in the finals themselves.
Finally, I’ll go over specific pieces of criteria that I’ve generally ignored. First, I haven’t included the idea that the MVP must come from one of the winning teams. While it’s generally true that the best player is correlated to the best team, that isn’t always the case. The best example I can give from 2018 is s1mple at StarLadder i-League S4. In that tournament, s1mple had the single best performance I’ve seen from any player ever. If you recall in the semifinals of that tournament, FaZe were up 1-0 on Na`Vi and could have potentially closed out the series. In the alternate universe where Na`Vi lose that series, I’d have not only nominated s1mple as the MVP, I’d have awarded it to him as well.
Another thing I’ve done is not take stats wholesale. The numbers in and of themselves won’t tell you definitively who was the best player unless someone breaks the bell curve (like s1mple did). So if you want to be exhaustive, you have to break down each game and figure out the value of the kills. As we know, in the game of CS:GO, we know that five kills does not always equate to five kills. Two players can have 20 kills each, but that can mean wildly different things. For now I’ll take SmithZz and apEX. If SmithZz is having an okay game, it’s fairly likely that his kills came from some duels, some AWP kills, some exits, some trades, and some ecos. If seen games where apEX also had twenty kills, but all twenty of them came on the T-side rifle rounds where he was literally breaking open entire sites by himself and ending the game 16-5 through his own brilliance. While the numbers are the same, the impact is so largely different they may as well be in different universes.
I’ll end this particular blog here as it’s gotten longer than I expected. In the next one, I’ll give a concrete example of my applied criteria in choosing an MVP candidate and in a later blog I’ll talk about decoding value and impact.