Within competitive CS:GO, we’re lucky enough to be involved in a game where there can be so many ways to measure a player’s performance, however, we haven’t quite reached the point of advanced analytics that mainstream sports have shown is possible. When you look at the NBA for example, in order to find the most advanced analytics you’d head over to the basketball reference and you can see thousands of stats on each player:
This is just an example of some of the stats an individual can see on players when they go on the website. For us in CS:GO, we have HLTV.org who have been able to provide more and more stats on players as the years go by. However, I think we’re not progressing fast enough when it comes to being capable of showing the stats or highlighting player’s performances. This is by no means saying that HLTV don’t do enough — they do — I just feel like we could be doing more. There are various sites that have been popping up, such as stayinpit.gg, SixteenZero.net. They’re producing fantastic content which I would suggest taking a look at if you’re interested in the stats-based/analytical content behind CS:GO.
The main point is, I feel like we should be capable of setting filters when looking at game overviews and looking back at games. Here are a few examples of things I feel like we could do with, when looking at player performances, or comparing players:
PerX Round Metric — this will work in a similar way to the per36 metric in Basketball. My suggestion for this would be to find the average amount of rounds played in a professional game, should be around 25, then use the average length as a way of measuring performance compared to measuring per round metrics.
PerX Buy Round Metric — this is similar to the PerX Round Metric, except that we now identify the average amount of buy rounds in a professional level game, this then needs to be filtered correctly, I would suggest this to be looking at equipment value, $4100 equipment value per player on T side, $4300 equipment value per player on CT side. By being able to establish this metric, we could identify how well players perform specifically in buy rounds alone, rather than in the game as a whole. It adds a whole new layer to measuring performance.
Close Game Performance Metric — in basketball, there’s a clutch rating for players. This is based off their performances in the last 5 minutes of the game, when the score is within 5. In CS, we use the word ‘clutch’ in different circumstances but ultimately the player would be considered ‘clutch’ all the same. When trying to make this metric work the same for CS, it’s somewhat difficult as there isn’t an allotted amount of time in CS whereas in Basketball there’s a set 4 quarters with 12 minutes in each quarter. One possibility, is to look at the overall game and consider rounds in this metric when it’s the following: team’s are within 3 rounds of one another, you or the opposition are within 3 rounds of winning. By being able to create a metric such as this, we could identify player’s who perform better when their team need them to, or when their backs are up against the wall.
Force Buy Specialists Metric — similar to the PerX Buy Round Metric, this would be a way of identifying how well players perform when their equipment is limited. My way of setting the criteria, would be to look at player’s performances whilst they are working with less than $2000 worth of equipment, but more than $950. I chose these values because of the following:
$950 — because that’s arguably the low end of a forcebuy. Example of purchases not considered to be a proper forcebuy: Deagle (no armour), p250 (no armour) + potentially grenades, etc.
$2000 — this is the higher end of a potential force buy, this could potentially rise to $2400 if I’m honest.
Creating metrics like this, allow us to sought out and separate the different levels of players in different circumstances. It also allows people to potentially become more wary of certain teams in certain scenarios.