When the MVP doesn't make sense

General Discussion
I feel like I am seeing more and more MVP awards that make zero sense.

Here's a great for instance out of the last game I was just in:

- I had 75% of my team's kills (the thing I was specificaly on the end game screen for)
- I did 62% of my team's nuke damage
- I "mostly" solo capped 4 bosses (IE boss was almost dead when a team member or two showed up and went "I'M HELPING!", where as I could have safely capped it completely myself).
- Finally, the end game UI said I did 72% more hero damage then the average Varian at my level.

MVP went to a player who's only posted stat for MVP was siege damage.

A while ago, someone on these forums stated that MVP should be more rare then it is; it shouldn't be an every game thing. But even at that, I feel like the person winning MVP rewards is feeling more and more wonkier the longer this system is live. It felt much more accurate earlier on.

I'm not talking about changes since where they added more cap rewards, I'm talking that it seemed more accurate on week 1 through week 3 of its release then it does now.

That being said, I don't think some categories carry enough weight. For instance, solo boss caps win games; picking up merc camps and pushing with them can offset the balance, but these sorts of things don't seem to affect who gets awarded MVP.

Other things that win games don't seem to be properly considered for MVP, such as showing up for and being a strong contributor in winning objectives.

Like the MMR system, I think the MVP system needs some more tweaking. Heck, it might even be a requirement to make the system ever-evolving to account for what wins based on each bracket of play.
The problem is that being top damage/siege/xp almost never influence the MVP, because what brings a lot more MVP points is takedowns and time spent death. Someone with 1 late game death but top siege, hero damage and xp is less likely to get the mvp than someone with the same number of takedowns and 0 death.
As Kenjin says, the MVP is determined using a very simple algorithm:

https://www.reddit.com/r/heroesofthestorm/comments/53av50/fyi_how_the_mvp_is_chosen/

This means that the MVP should just be seen as something which highlights a few figures after the game. But by no means it is representative of someone's skill or performance.
12/01/2016 02:02 AMPosted by Lighthammer
I feel like I am seeing more and more MVP awards that make zero sense.
I've never seen one that makes -zero- sense. Especially if you know how it's calculated. It's just not always the person I would have picked.

12/01/2016 02:02 AMPosted by Lighthammer
- I had
12/01/2016 02:02 AMPosted by Lighthammer
- I did
12/01/2016 02:02 AMPosted by Lighthammer
- I "mostly"
Well, here's your problem. You're playing a team game and thinking only in terms of yourself. One of the things I enjoy about the MVP screen is that it quite often manages to find that one player who may not have had the most damage, may not have had the most siege, but was actually there working with the team for almost every single kill.
If you have 75% of teams kills you get 1 point for it. Nukes don't count. Solo capping bosses is risky AF. So based on this you got 1 category where you were top, it's pretty rare to see person like that get MVP...
Do you have the score screen by any chance? Whenever someone complains about not being MVP* and they do post stats, there's usually at least one other player with comparable stats and a slightly better takedown/death ratio.

*And it is always 'Why am I not MVP', never 'why isn't player X MVP'.
Actually the MVP algorithm isnt that simple.

Every action has to be translated in one specific "MVP calculation exchange value" (because otherwise you cant compare 3 merc camps to 100.000 heroe damage). So you need to set specifc amounts of this mvp calculation exchange value for every specific action. Then you add them up, for every single action that is currently taken into account and you substract this by a specific amount for every single death.

I have already seen players getting MVP that had a few more deaths but way more impact. I have also seen players with slightly less impact but way less deaths. So I wouldnt directly say that the whole system isnt working.

Right now deaths decrease your chances on getting MVP by a lot but thats completely fine. Players need to understand that a death is a huge - HUGE - drawback. You might lose soak, you might lose an objective, you lose map presence, your enemys can play more aggressively because they have an advantage, its harder for you to have some gank rotations without defending a push and many more stuff like that.

So its not impossible to get MVP with extremely good stats and a few more deaths but in exchange deaths are realy punished. And its not about gold, MMR, achievements or stuff like that anyway. Right now its like a little hint. So if you are Valla for example and you always top damage but die 7 times per game ... never getting MVP ... you seriously should think about that. Wouldnt your impact have been greater if you were less often death ? And didnt your "huge damage numbers" come from bad decisions anyway (thats the irony many players dont understand - that engaging but dying all the time automatically leads to much higher damage numbers than only taking good fights without dying).

But by no means it is representative of someone's skill or performance.


Wouldnt say this. I wouldnt say that the MVP system is 100% accurate but to say it doesnt reflect skill or performance at all is wrong aswell. MVP Rich - one of the best players in the world - has a MVP ratio of 26%. The average ratio would be 10%. So its pretty obvious that skill has something to do with getting MVP.

Sidenote: he also has 71% winrate. How can this be ? Isnt his winrate forced to be exactly 50% ?! xD
12/01/2016 02:55 AMPosted by LowBob


But by no means it is representative of someone's skill or performance.


Wouldnt say this. I wouldnt say that the MVP system is 100% accurate but to say it doesnt reflect skill or performance at all is wrong aswell. MVP Rich - one of the best players in the world - has a MVP ratio of 26%. The average ratio would be 10%. So its pretty obvious that skill has something to do with getting MVP


Actually we agree, so maybe I should rephrase my opinion: the fact that one got MVP at the end of the game does not necessarily mean he's been the best player of this single game, neither that he is the highest skill player in this single game. But yes, on the long run, someone who gets MVP in more than 10 % of his games is mathematically better than average.

12/01/2016 02:55 AMPosted by LowBob


Sidenote: he also has 71% winrate. How can this be ? Isnt his winrate forced to be exactly 50% ?! xD


Agreed. Though I haven't seen anybody on this thread saying the opposite.
Actually we agree, so maybe I should rephrase my opinion: the fact that one got MVP at the end of the game does not necessarily mean he's been the best player of this single game, neither that he is the highest skill player in this single game. But yes, on the long run, someone who gets MVP in more than 10 % of his games is mathematically better than average.


Pretty much this.
In the end we all have to keep in mind that its some algorithm calculating the MVP. So its more like a hint and people should use it like that. Its also a title for average impact. So for example someone can be MVP, have the most overall impact but maybe someone else made THE ONE big lategame play that made a comeback possible or something like that. Things get complicated in such scenarios because its pretty subjective who was the better player - the dude with permanent solid impact or the dude with the most important single situation ?
I think its fairly objective that the MVP should be the person who made the biggest difference in winning the game.

A Valuable Player is straight up someone who highly contributes to your success in a match, the MOST Valuable Player is the guy who made contributions far exceeding.

Thats why I go back to a previous poster and say that not all games necessarily SHOULD have an MVP.

But on the same token, if the things you are doing in that match are the things that are winning the game, the MVP title SHOULD go to you.

I can't tell you how many games where MY TEAM recognized me as the MVP with their votes while the person given the MVP was just kind of "there".
I don't think getting takedowns and not dying necessarly means you had the biggest impact.

Dying 4 times early game is a big deal, yet in the MVP system it is as much punished as a late game death. Because what's important is time spent dead, not number of deaths.
And even if a full team wipe is far more impectful late game, one signle death isn't.

Also, that means if you're solo laning and too far from kills to get the assists, you are guaranteed not to get the MVP.

I think the MVP should give more important to statistics, even when you're not top, and give less importance to takedowns/deaths, while also counting number of death and not time spend dead.
The thing I'm constantly weighing in my head when it comes to MVP status is impact.

If there was a fixed "If you cap a lot of merc camps, you could get MVP" feels like a terribly bad measure largely because those merc camp caps are only useful if done at the right time for the right reason. Solely capping camps isn't a good strategy and you shouldn't be rewarded for JUST THAT.

The same thing follows with many of the main categories you might use. Getting a lot of kills doesn't mean much if you constantly over extend and constantly break your team's line in doing so.

The measures used currently don't necessarily dictate the player who had the most impact nor it shouldn't encourage you to do things just for the recognition.

I feel like the new UI post game stats that tells you how you performed in relationship to other people at your level on that hero needs to weigh into MVP.

Another one that amazes me that you're not recognized for is the whole solo or near solo boss caps. Those can win games, especially in the mid to lower levels. Just because you're on a hero capable of it, doesn't mean you can do it; its something that takes a lot of skill; something I think needs deeper recognition when you accomplish and recognition when you botch it.

Related-ish: I think with the whole Political Correctness / Special Snow Flake period beginning to come to an end in US Society; I think the game DOES need to tell you when you screwed up too. This can be done in a way that doesn't elicit toxicity, but does tell you to "get good". Too many people in this game have developed some really bad habits and there certainly needs to be some sense of reality checks.
12/01/2016 02:19 AMPosted by PlatinumPOS
One of the things I enjoy about the MVP screen is that it quite often manages to find that one player who may not have had the most damage, may not have had the most siege, but was actually there working with the team for almost every single kill.
Being present for the most kills doesn't make them the best player for that game. Quarterbacks do not automatically win the MVP award at the end of each game just for showing up in the most plays, they win it for outstanding individual performance. And the reality of the situation is that MVP is an award which the computer can never accurately calculate, because there are simply too many variables that go into what makes a player "MVP" that the game simply doesn't think are important.

An E.T.C. who lands multiple fight-winning Mosh Pits should get consideration for that, but stuns/silences/roots are ignored in the algorithm. Kills are counted equally to assists, meaning the player who secures 12 kills is ranked equally to the player who is present for 12 assists. There's no consideration for how much higher a stat is from the rest of the team or from the average of that hero, so a player can compile incredible hero damage that helped win every team fight, or lay down critical siege pressure that kept your team ahead in structures, and still get no more than one point for those efforts. And these are just some of the many, many flaws with the overly simplistic algorithm in use for the MVP award. That's precisely why it needs to be removed and replaced with one that lets you vote for who the MVP award should go to from any of the ten players present in that game. You're more likely to end up with the actual MVP if you let the players decide, rather than trying to get a computer to pick the best player and doing it poorly.
12/01/2016 04:35 AMPosted by CriticKitten
12/01/2016 02:19 AMPosted by PlatinumPOS
One of the things I enjoy about the MVP screen is that it quite often manages to find that one player who may not have had the most damage, may not have had the most siege, but was actually there working with the team for almost every single kill.
Being present for the most kills doesn't make them the best player for that game.
But it does make them the biggest team player, which is a great place to start.

12/01/2016 04:35 AMPosted by CriticKitten
Quarterbacks do not automatically win the MVP award at the end of each game just for showing up in the most plays, they win it for outstanding individual performance.
This was probably not the best example you could have used, as Quarterbacks are typically considered literally the "most valuable player" on a team by default. Take a look at NFL salaries by position.

12/01/2016 04:35 AMPosted by CriticKitten
An E.T.C. who lands multiple fight-winning Mosh Pits should get consideration for that
I agree completely.

12/01/2016 04:35 AMPosted by CriticKitten
but stuns/silences/roots are ignored in the algorithm. Kills are counted equally to assists, meaning the player who secures 12 kills is ranked equally to the player who is present for 12 assists. There's no consideration for how much higher a stat is from the rest of the team or from the average of that hero, so a player can compile incredible hero damage that helped win every team fight, or lay down critical siege pressure that kept your team ahead in structures, and still get no more than one point for those efforts. And these are just some of the many, many flaws with the overly simplistic algorithm in use for the MVP award.
Do we know how much of this is correct? Because as far as I'm aware (and correct me if I'm wrong), nobody has access to the specific algorithm used to calculate the MVP. Do we really know that kills and assists are considered equal, or is that just conjecture? Do we know that it doesn't take stuns, silences, and roots into account? Or is that also conjecture?

I'm asking because in my experience the MVP does seem to take all of these things into account. I've seen multiple (tanks especially) players who had a large part in game's success without having the most kills be rewarded MVP. I play Alarak and am frequently labeled "Silencer" after the game with how many total seconds of silence I inflicted. It seems these things are indeed taken into account, so I'm interested if there is any definitive information on this. From what I understand, the most people have been able to do is guess.

12/01/2016 04:35 AMPosted by CriticKitten
You're more likely to end up with the actual MVP if you let the players decide, rather than trying to get a computer to pick the best player and doing it poorly.
Again, correct me if I'm wrong, but I believe this is the entire purpose of the voting system on the MVP screen. It gives the players themselves a chance to confirm/deny the computer's pick.

Join the Conversation

Return to Forum