Forum

How would you measure class perfomance in a game?

Created 10th December 2014 @ 09:13

Add A Reply Pages: 1 2 Next »

Kengur

I mean something that could be formulated in math and could be comparable for all classes… This could also change depending on map type – e.g. on pl scout should be on cart half the map and he’s getting little kills.

For example:
Assume meaningful scale between 0 and say 1000. We try to assess heavy. Lets say we want to look at damage, spies killed, medic defence and time alive.

Heavy KPI = (total map time / total time alive ) * ( ( damage done – enemy heavy damage done ) / 10 ) + spies killed * 10 – medic deaths * 100 )

Same for scout… we want him to kill or pop enemy med and cap points

Scout CP KPI = enemy med killed * 300 + enemy med poped after some reasonable damage by you * 100 + points capped * 150


Last edited by Kengur,

Bloodis

(Legend)
(๑╹ω╹๑ )

idk exactly but in any map the more airshots you have as any class, the better you are

the end?

doks

You can`t measure someones performance with an equation. Not even by logs. There is a lot of work being done, that won`t show up on logs or anywhere else. How would you rate someones game sense or calls being made ? How will you rate someone calling out a spy behind your medic ? Who gets profit ? The one killing a called out spy or the one who noticed him ? How will you rate a medic – saving a soldier from falling to his doom with an arrow ? or heals ? etc etc…


Last edited by doks,

Gentleman Jon

Defining stereotypical roles for classes then rating their stats based on it punishes people who play a class in an unorthodox but successful way.

Kengur

Quoted from doks

You can`t measure someones performance with an equation. Not even by logs. There is a lot of work being done, that won`t show up on logs or anywhere else. How would you rate someones game sense or calls being made ? How will you rate someone calling out a spy behind your medic ? Who gets profit ? The one killing a called out spy or the one who noticed him ? How will you rate a medic – saving a soldier from falling to his doom with an arrow ? or heals ? etc etc…

Well in fact you can measure all of this from logs, except calls. But does it really matter if the solli dies or not?

Collaide

The way easiest way would be splitting it up into 9 sorts of ranks, one for each class, where your statistics mattered, but also puts an emphasis on the stats of you opponents, the game over-all, the classes your opponents and teammates have & are using etc.

The reason behind I think there should be 9 sorts of ranks (and then that is divided to a lot of subgroups for that class) is because you can’t really judge a class performance based on an entirely different class.

If you are talking about the potential in-game competitive system I think this will be the best.
Maybe one could ban 2 classes just like in Arena: Respawn or something idk.

Kengur

Quoted from Gentleman Jon

Defining stereotypical roles for classes then rating their stats based on it punishes people who play a class in an unorthodox but successful way.

Well I am talking fuzzy logic here. And I definitely need more oppinions on what are the KEY objectives of you class and how would you personally judge if you did “bad” or “good” in a single game. If you try to analyze it and get some connections out between pure numbers and those human readable assessments, that would be helpful.

Kengur

Quoted from Collaide

If you are talking about the potential in-game competitive system I think this will be the best.

Yes, the idea is to use this as a basis for a rating system. But for it really to be any useful we need as much info as we can get away with, including individual game performance.

UGC for example as a single rating which goes up or down fixed ammount based on your team loose or win. This is totally useless to assess individual player rating. So individual player rating is wrong, so you can’t for example assess team rating based on a collection of it’s players ratings etc.

fraac

JOHN
CENATION

Isn’t this much harder if you’re only considering players within teams, say from pcw logs, rather than from pickup logs? How could you tell the difference between an uncreative roamer and a good playmaking roamer on a weaker team, if their individual stats looked the same and the teams’ results looked the same?

Gentleman Jon

Quoted from Kengur

Well I am talking fuzzy logic here. And I definitely need more oppinions on what are the KEY objectives of you class and how would you personally judge if you did “bad” or “good” in a single game. If you try to analyze it and get some connections out between pure numbers and those human readable assessments, that would be helpful.

If the statistics aren’t linked to the game result then you’re just asking people to mark their own homework and declare themselves a success.

In the example above why would you circumscribe the scouts role so much? If a scout kills literally everyone except the medic you’re giving him no credit despite the fact he just won the game.

If you ask 10 people on this you will get 10 different answers, each one based on their own perception and prejudices and preferences.

The only way to get an objective picture is to measure their importance in relation to the result.


Last edited by Gentleman Jon,

smziii

(Legendary Ratehacks)
SVIFT

Teamresult > personal fraghunt

Collaide

Quoted from fraac

Isn’t this much harder if you’re only considering players within teams, say from pcw logs, rather than from pickup logs? How could you tell the difference between an uncreative roamer and a good playmaking roamer on a weaker team, if their individual stats looked the same and the teams’ results looked the same?

Having two different queues maybe? Solo/duo & team? 8)

Kaelan Frey

The easiest (but probably not very efficient) possible way for DM classes to be measured in performance is in how effectively they can gain territory for their team during the game (when on the offensive, of course… in defense it’s the preservation of territory that matters)

For example, it would be unfair for a Pyro in 6v6 to be judged by their kills/damage, because his entire purpose is area denial – but their airblasting ability is unique, which can halt a lot of pushes very effectively and sometimes even pin down certain classes for easy picks which he won’t do, the Scout or the Soldier is quicker at doing that.

I’m not sure how it works on Highlander, but the inherent concept should be the same.


Last edited by Kaelan Frey,

ash

(Legend)
:D

Quoted from Kaelan Frey

We could define certain areas of a map as ‘controlled’ by one team whenever they hold a key location for a certain amount of time.

Let’s call it capture points.

stuntz

you cant have a formula for that kind of things, imagine this scenario you have a scout trolling around 4 people while preventing the point to be captured he died and did jack shit damage but sucessfully stalled 4 people from capping unting his team arrives, while he didnt do any damage he was still usefull in that situation. Its like having a spy cap in every round, while you dont do damage youre still being usefull

xeerszz

Add A Reply Pages: 1 2 Next »