Valorant impact value: giving credit where credit is due
It is indisputable that current Valorant stats are shallow, totally duelist-oriented and do not reflect the true impact of certain actions within the match. Let’s say you are playing Split defense and you guys just won the pistol round against attackers. On the second round, some of your hungry team mates randomly decides to overextend mid and he ends up taking multiple gunfights at once. Roughly speaking: can it be an impactful play? Let’s dive into it
- Outcome A: he gets 3 kills and falls back into cover
- Outcome B: he gets the same 3 kills, but dies and gives his Spectre to his enemies
- Outcome C: he does not get a single kill, dies and and gives his Spectre to his enemies
Combinatorily speaking, going for multiple gunfights at once is a task classified as high risk (there is a chance of gifting your spectre to enemies) low reward (regardless of what you do, your team should always be winning that round). That being said, the duty of winning the round is never in the shoulders of the team that lost the pistol. Because of this, each kill the attackers get brings way more value to the table than each kill the defenders get.
That theory sounds cool on paper, but does it really works like that on real matches? Let’s look at one example and see if things goes as we expected :)
As we can see in the table, defender kills (and multi-kills) increases little to nothing the chances of defenders winning an already favourable round, while any kill the attacker team gets flips the odds by a huge margin.
Although both MkaeL and dapr got 3 kills, just one of them was supposed to win that round, the other was not. This leads to a 109.99 delta impact for MkaeL and only -0.34 for dapr. Cool, no?
Let’s take a look at how ACS treated the same situation
As you probably already expected, ACS treated both situations as if they were equally impactful, which is far away from being true. I mean, what ACS fails to achieve is capturing the nuances of each situation. Not every multi-kill is the same. Not every gunfight tells the same story. Not every round has the same odds.
What if you got a multi-kill while holding your operator on an empty bombsite? According to ACS you have done one of the best plays you could do to your team. But at the end of the day you did not impact the odds of winning that round, since the bomb was already going to explode with or without your flashy multi-kill.
By saying this i’m not trying to bash on that metric and I totally understand that Riot wanted to have at least a placeholder. After all, having a score system is better than having no score system, isn’t?
Who knows in the future we can have keep improving on this and implement better metrics for non-duelist agents? Let’s give Caesar what belongs to Caesar :)
Final scoreboard above. We all know sage players have a hard time on Icebox. Look at how SicK went from bottom to 2nd place, climbing up +3 positions! As much as he had the worst KDA in the whole team (11/14), his 11 kills were impactful. After all, that is what should matter in a Valorant match.
