Adjust Hero Power Calculation

Why does it matter:

Hero power is used in Alliance War and Raid Tournament matchmaking. The players/alliances who owned heroes whose actual power are below what is displayed will be at disadvantage and vice versa. Do you really believe that Costumed Elkanen worth more (798) than Frigg (786)?

The current system:

Based in this post, the current system of power calculation is:

0.35 × Attack (excluding from emblem) + 0.28 × Defense (excluding from emblem) + 0.14 × HP (excluding from emblem) + 5 × (Skill Level - 1) + Star Base Power + 5 × Emblem Node

Where Star Base Power is:
1* 》 0
2* 》 10
3* 》 30
4* 》 50
5* 》 90

As higher star generally have better skill than lower star, we can assume that Star Base Power is actually tied to the Skill’s worth at level 1.

What the current system fail to take into account:

  • Mana Regeneration Bonus from Costume
  • Family Bonus
  • Realm Bonus
  • Passive Ability
  • Skill Worth Difference (most game changing).

How to improve:

Aside from adding minor points for the things that obviously missing, SG should take into account skill’s worth. Even among heroes with the same star and speed, there are skills that are bad, average, good, and even OP. However all of them had the same power points of 90 + 5 × (Skill Level -1) for 5*. Imo, heroes with blantantly stronger skill should worth more points.

I’m not quite sure this is right.

Based on this model, Sartana should have:

  • 90 Power as a 5*
  • 0.54×694=374.76 Power from her attack stat
  • 0.45×700=315 Power from her defense stat
  • 0.27×1382=373.14 Power from her health stat

This adds up to a total of 1152.9 Power. Or 1152 when rounded down. This is significantly higher than her actual 757.

1 Like

Not exactly sure how would you quantify the “worth” of a skill… who’s skill gets a better rating? The same skill has different field value depending on the moment it is used. Some skills are of great value in the biggining of a fight, when all your team is up and running, some have greater value towards the end of a fight, like Owl’s insane damage if he’s the sole survivor and will take out all the remaining enemies with one hit. Of course, most would say Bera’s skill is far superior, but when only one of your heroes is still alive and the battle is undecided, I’s definitely wish I had Owl on the field rather than Bera… this is all subjective and relative, can’t really place a fix value on the “worth” of a skill… of course, you’re right about the special skill being more important than hero stats but quantifying it could ever only be subjective, never objective…

2 Likes

There is certainly an argument that the context of the skill plays a role, but I believe a subjective consensus on the skill could be found.

Currently for a newer player, all they have to go on are TPs, but you could argue that SGG could advertise heroes and intended usages far better. A simple guide (even if it’s locked in the TP calculation) or something else to help players decide who to level would only serve to better the skill and understanding of the playerbase. A simple rating of Niche/Great/Outstanding would certainly benefit the players even if not locked away in TP. Many players ascend the wrong hero without understanding how/when to use the hero.

The other benefit is that SGG would also consider the relative strengths of the skills they release. It’s a crime that there are heroes who still deserve better handling of their skills (Noor, Inari, Myztero etc etc.).

I’m sure while it may not be objectively clear that Bera is Better than G. Owl. We could easily come to consensus that Bera’s skill is better (Outstanding) more of the time than G. Owls (Niche) Even if you you were the greatest proponent of G. Owl.

Very well articulated @yelnats_24! This could make for a decent player guide.

I’m partially okay with power creep because it means my season one heroes aren’t expected to match up with newer ones.

Theoretically if matchmaking were good then power wouldn’t matter because we could just go off of Elo or some equivalent.

Perhaps power should be adjusted by usage? We know about half of top 100 defenses use Frigg and Odin. So perhaps that means they get an extra 20 points if power based on empirical evidence they’re good.

I don’t think that’s true. To be able to ascend a hero, you need to get the AM’s needed for it and those mats take time to gather. I’d say, 6 month it’s a safe bet for the MINIMUM amount of time needed by an average player to get the mats for the first ascension. But I think it takes more than 6 month for most players. If you still pay attention to a hero’s power index after 6 month, then you really have a big problem. Assessing a hero by their power index is the worst mistake anyone can make. You need to focus on battlefield action, your own experience while facing that particular hero to determine if it’s worth the mats, not on some numbers next to it’s picture…

1 Like

I think the other thing is that certain special skills are more important/ stronger depending on the current meta. When everything was heavy DoT for example the Hatter wasnt a viable option and lots of people wouldnt have considered levelling the likes of Grimble before minionpalooza with Freya, Bera, Krampus etc

It’s hard to give a power value for skill type when 6months down the line it could become niche as the meta shifts.

I posted the wrong version, should have been 0.35, 0.28, and 0.14 instead of 0.54, 0.45, and 0.27. Thanks for the correction.

2 Likes

This is a good idea…
You have my vote :slight_smile:

1 Like

While I absolutely agree that this is a fact, SGG will never ever ever officially admit that some skills are bad and others OP.

Here are the problems with that:

  1. First and foremost, assigning a “worth” class to a hero would directly imply whether or not they’re worth to spend to summon.
    Right now, by default, for most players, any new heroes look shiny and worthwhile. Assigning heroes “bad/average” designation will make players less likely to want them.

  2. Thinking the SGG official dev team can correctly estimate skill class on hero release would be thinking too highly, given their track record.

  3. This will entirely fail to account for context, niche heroes and counters, since a lot of heroes:

  • Are amazing on offense, bad on defense
  • Are amazing against minion comps, useless against comps without minions
  • Entirely shut down certain strong/meta heroes, but will fall out of grace when the meta inevitably changes

I do agree that the current hero power values are misleading, but there’s no way to make them not-misleading while preserving the charade that all heroes are worth summoning.

1 Like

Finally someone else into the “how stupid the hero power metric is”.
Harsh? Maybe, but it’s so stupid that

Ditto. Obviously… that’s exactly the metric used by SG for matchmaking so draw your conclusions.

While it is unlikely to ever have a good general skill performance index, it’s definitely possible to have a better one.
Just as an example, we could rate snipers on the % of dam dealt weighed on their attack stat. Secondary effects to add.

While it could be hard and prove ineffective, it’s even harder to sort out something less effective than what we have so we really don’t have any reason not to try: decent information on a limited number of instances is better than no information whatsoever.

But we don’t need to face the task by defining hero specialties and then trying to define a performance index for every specialty (although that could be a sound approach).

We could solve the problem by using hero academy and letting players select a preference list on the hero they would like to show up thru retrains. You wouldn’t even need to disclose percentages, nor to make it a permanent feature.
That would reflect the value the community assigns to every hero.
By inspecting the lists, weighing 'em on the number of heroes in the roster and on players prestige/reputation you could come up with the hero’s worth.

It could also be run as a poll in the community, without any need to code it into the game.

But the general rule applies: when improving something you don’t need to make it perfect, you need to make it better.
And it doesn’t look that much of a challenge here.

1 Like

What is 0.54 , 0.47 and 0.27 ?