On Hero Mastery

Moderator note: Moved this discussion from the thread on Excoundrel’s 4.1 preview. ~HS

I think the API land scape has been rocky for the past few months ever since they dumped rotating game modes into the game.

While Mastery is still barely useful for anything other than enabling assurance that a player might be good at the hero they’ve pick to their teammates during loading screen… I have to say it is a subtle but smart move by SEMC.

It gives slight meaning for players when they play modes that are not currently tracked by API sites. It’s nothing much, but it’s… something.

My hopes is that SEMC sorts out their API issues so sites like VGPro have an easier time to update.
VainAura at least have San Feng stats for high tiers, but that’s not really useful.

Another hope is that they up the Mastery limits.

I imagine that they can just add on numbers to the system, and the stars change colours the higher you go after hitting five stars of a set limit.
Say, you have 5 star mastery on a hero now. They up the limit and the next star is bronze (your other 4 slots will remain white so people can see that you have actually hit 5 stars white and such. Making them blank would be dubious imo).

I do wonder if this new stuff they are saying they’re making have anything to do with Mastery stars.

Mastery stars mean nothing



To you, maybe. To others, depends on them.

Mastery Stars to me is simply a way to show others that a player has played this hero a lot and can be somewhat reliable.

I say somewhat reliable because Mastery points are afforded at the same rate regardless of win or loss, which I think is a mistake.

Losing matches should only give half of what winning matches give.

I still reserve the idea that it could be better, and the visionary in me sees many ways it can be extrapolated and improved upon.

That said, SEMC has a track record of crushing dreams and hopes, so I rather just sit and watch how it goes.


This doesn’t mean someone is good with that hero, I found a t4 with 5 stars with Ringo that didn’t know how to even build Ringo. Master points only mean something in really high elo (VG silver and gold) IMO


To have a meaning, they should tweak the number of stars + how many games you should have with a given hero to add one, AND to reset them every season. Otherwise after a year from now, everyone will have a lot of stars on a lot of heroes and this will totally remove the value/info they carry. :slight_smile:

Resetting them isn’t the answer. To have value as a marker of “mastery” the stars have to be related to performance: in other words, you have to be able to LOSE them as well as gain them based on how you play. Ideally, they should measure something other than simple wins and losses – even KDA would be better than just how many times you played a hero. And if you fall below a certain target – by losing a match or if your KDA is 20% below expected for that role in that type of match – you lose points.

But that’s complicated, and it requires data that SEMC doesn’t provide to the API, so I have no clue whether they even track it. (Based on what I know of how their playtesting works, I’d say that they don’t actually have that sort of data themselves.) When VainSocial was alive, I created what I thought was a useful mastery system which got a lot of positive feedback from high level players. However, that required the massive database that VainSocial had built just to analyze tens of thousands of matches. There’s no doubt in my mind that SEMC doesn’t do anything like that.

So they just did the easy thing, which was to give you 40 or 80 points each time you play a match. :roll_eyes:



A simple decay on the points would suffice and eventually when or if Vg gets over 80 heros, we can revert the decaying to back to un-reseting or un-decaying to be able to maintain masteries.

Similar to LoL which as a fine and working system.

I disagree, hero mastery is something individual and SHOULDN’T be able to lose just by losing a match.

You have a troll in you team that eventually leads to a loss while you manage to preserve a good kda. But in the end you lose points?


I should have worded that better, because obviously it should depend on actual performance, as I used in my system for Andi. I only mentioned that because I can see SEMC taking the easy way and having you lose stars because of a loss. After all, that’s how VST works, right?

1 Like

There’s actually a much better way to do that, which is to give more recent performance more weight in calculating mastery score. That way, it more accurately reflects the level a player is currently likely to play with a given hero. For players that haven’t played recently, you could have some sort of visual indicator (such as dimming the stars) that they haven’t played that hero in a while.


KDA should not be measured. That is way less accurate than win %

In some ways, yes. Sure, you can have say 9/3/11 and be playing poorly because you’re not helping your team take objectives etc. But on the other hand, you can have the same score, be playing absolutely perfectly, and lose the match cause your teammates are rubbish. Both have their advantages and disadvantages when calculating mastery.

Sure you can lose one game because of bad teammates, but your win rate is the best possible reflection of how much you contribute to a team on average. If you have a 40% win rate but high kda you probably are not very good.

You can also have a bad kda and win which would be equally likely, and it’s impossible to measure captains with KDA or any other metric. Additionally certain roles simply tend to have better kda. bot and mid lane are often protected and supported by the team given more gold and safety making it more likely they will have a high KDA regardless of how good they are. You aren’t thinking about averages or the impact this would have on non carry players. On average win rate and by extension rank is a far better representation of skill than kda.

Mastery has no strategic meaning as it doesn’t show up during Hero Lobby or Draft, nor does it mean anything as it’s implemented SO late.

To me, a Badge system works better, you could have multiple badges unique to each heroes that measure the level of skills- not the frequency of play. Example:

“Insert Star Here”- (0/500) Kill enemies with Heliogenesis
“Black Hole”- (0/200) Stun 2 or more enemies with Core Colapse
“Light show”- (0/100) Break enemies out of Stealth with Julia’s Light
“I rEqUIrE hEaLinG”- (0/100) Heal ally who’s health is below 10%
“Judgement!”- (0/300) Kill enemies with an empowered attack
“Green Energy”- (0/10000) Regain 10000 energy using Perk.

Badges must be completed within a season. If the season ends without the goal being reached, the progress will resets. Earned Badges are pernament.

The system can show how proficient a player is with different play styles of the same hero. If you are a CP Adagio, you’ll more likely to earn “Judgement” than other badges. If your team is playing against stealthy enemies, it’s less worrying to have a Celeste with a “Light Show” Badge.


I like this idea a lot, but would probably require quite some coding, and considering how buggy the game is as is I’m not sure I trust SEMC to do this.

Also one would have to be very careful as to how the badges are earned, I can imagine people not quite apt enough attempting to earn them and playing stupidly as a result, tilting both them and their allies.
For example I can imagine an Adagio chasing enemies under turrets just to gain energy, or holding back on healing allies until they’re at low health and accidentally letting them die.

1 Like

The problem is that SEMC doesn’t actually collect much if any of this information, so as @MacAulay says, it would require significant changes/additions to the code to implement any such system – which I totally agree would be far more meaningful than the rather stupidly simple system they’ve thrown together with minimal effort implemented in 4.0.

It’s tragic, because VG really has a ton of potential to outshine anything out there, but its developers keep squandering it.


Well it’s my first time someone saying I was playing my hero bad while I was killing enemies left and right , then it all made sense when he said he has 5 stars on that hero , I played that hero since the release and Iam playing him in different accounts , Iam actively until now playing the hero not even taking a break and getting back to the game to judge me.

I thought the stars means the player familiar with the hero , not expert , because it based on nothing , play more you gain more stars.

Who are you replying to?

I got the notification, so I will answer you: familiar with ‘X’ hero doesn’t mean good with it, or competent, I can play a lot with Lyra and still be the worst Lyra you can ever find. Practice is not always progression.

Many people ‘practice’ some heroes in casual, but that only helps you to learn the basis of that hero, so at certain point, playing casual doesn’t make you better with it, and if we are talking about ranked, if that practice is not translated in achieving higher ranks, then your stars means nothing because you are not improving your skill or knowledge about that hero.

1 Like

It was a general reply about the hero mastery , I agree with you , hero mastery does nothing , to learn heroes you need to watch good players playing them and take notes then practice , and add to that knowing the other heroes too so you can outplay them with your hero , so we can tell everyone has 5 stars on certain hero that you are familiar with it in this meta good luck learning them again in the next meta , and when they turn it to a potatoe your 5stars gone for nothing .

That’s actually a GREAT point.
But if I was SEMC, I would make sure that players are not awarded for those type of behaviors. So if your team lost OR you get more than 4 reports for eg, then no progress will be made towards the badges.

1 Like

That’s a good idea, but I’m not really sure it would stop people from doing it, especially at lower tiers. Sure, they wouldn’t get the badges, but they’d still try.