Wednesday, March 30, 2011

Automatically Measuring Contribution

Lately, more and more games with group activities have been trying to hand out rewards based on the individual's contribution to group's success. The latest example is rewards for defeating rifts in RIFT. The problem with all these systems is that contribution is very hard to measure automatically. It's bounded by role, by level, and by strategy. Trying to say that Joe contributed more than Randolph, and thus deserves more reward, is a very hard problem.

It's even harder once players start trying to game the system. For example, RIFT apparently used the rate at which you used abilities to determine contribution. Which sounds pretty reasonable at first. If you constantly attacking, or constantly healing, you're probably contributing a lot. But players then started spamming buffs, or spamming abilities like "Track Minerals" to get to the top of the contribution meter. Obviously, Trion will fix this, but it illustrates how hard measuring contribution is.

I wonder if the better tactic would be to change the question. Instead of How much did the player contribute?, maybe the game should just ask: Did the player contribute? Instead of trying to quantify contribution, just make the question binary. And then just more or less equally reward everyone who did contribute.

I think the binary question is much easier to answer. For a game without defense, I would just use the damage done, healing or damage taken meters. If you are above a certain threshold of the top value on at least one meter, you probably did contribute to the group's success. It would be a low threshold, like say 20%. If you did more than 20% of the top damage dealer, you'd count as a contributing dps.

That's enough to give you a shot at the rewards, in my opinion. It may not be perfectly fair, but it's "fair enough", and all the other systems seem be able to be exploited more easily.

Now defense is much harder to measure. Let's say you're defending the mine in Arathi Basin, and no enemies attack you. You're still contributing to the group's success, but you don't show up on any of the meters. Maybe one solution would have the defending point give out a buff in a certain radius, and then have a meter count the time spent with that buff. A "time defending" meter. And then use the same threshold system as for the other meters.

So to conclude, quantifying contribution automatically is hard and vulnerable to exploitation. Reducing the question to a binary "Did the player contribute?" might be easier and fair enough to satisfy the player base.

12 comments:

Nils said...

One of the most powerful tools are the players. For example, you could have them make groups.

Players would use the human brain and its extraordinarily evoled justice-mechanism to answer the question of who has contributed how much ans deserves what reward.

Of course, we know the problems with that approach, too. :)

Hobart said...

I think there are too many variables to measuring contribution to have it done by algorythms. What if

-a player sucks at his task but tries never the less? Say "you didn't contribute (=you suck)" to his face? VERY discouraging..
-a player defends a point of strategic value that the game company didn't think of (not defend the flag itself but a bottleneck path nearby)
-you are a healer but can't contribute effectively because you are in constant stunlock situations

Whenever contribution is measured automatically, there are bots and players exploiting the system. It always ends up unfair. Also, you are killing all "Think out of the box" strategies (which you kind of like, no?)

Kring said...

Most raids in WoW use DKP or /random to distribute the loot, even random /trade raids. The /random system does not consider contribution and gives everyone an equal shot who was present. And DKP only consideres your time investment but not your actual contribution.

Rewarding real contribution is not important, otherwise the player would have invented it long time ago.

Anonymous said...

It should be noted that Rift patched in a change to the contribution. Mainly they removed it. You now get loot based on how many stages you were there for.

Babel said...

Anon,

I must have missed that in the patch notes, though there has to be something more than that. If I went to a rift and did nothing I wouldn't get anything. I'd hate to see players just hanging outside of rifts to ge the rewards and only having 1-2 people attacking while a large group is afk.

Brunnstag said...

Honestly, I think having computer try to measure all these different variables will never work appropriately. There's just always going to be someone who does something in such a way that the algorithms will miss.

Thankfully I've never had to run in a raid where I felt other people were not contributing. If your there and your trying, your still contributing something! The fact that your output is not as some of the others is sometimes even an indication that you need the gear more than some of the other players!

Hobart said...


"Rewarding real contribution is not important, otherwise the player would have invented it long time ago."


Try joining a /trade group and not do your job, see how much loot there is to grab :)

Usually the common understanding in a raid is "unless you fuck up badly, you have an equal chance for loot" which means everybody is measuring your contribution and deciding you are either able or not. This happens all the time, its just not set in stone (which is good because that would lead to exploitation).

Kring said...

> Usually the common understanding
> in a raid is "unless you fuck up
> badly, you have an equal chance
> for loot"

That's basically what I've meant. A possible solution would be:

less than 10k dps: no loot
10k - 11k: /random 70
11k - 12k: /random 80
12k - 13k: /random 90
13k+: /random 100

Stuff like that would be really easy to implement and I've never seen or heart of a raid that does that (besides the "no loot below x dps" rule).

Chris said...

Good post and quite interesting, but please stop calling the game "Rifts." That's not a name, it's a feature. The game is called "RIFT" -- capitalized because that's how it's marketed. It would cut down on confusing when referring to rifts in RIFT.

Rohan said...

That's a good call, Chris. Will edit the post.

Hobart said...

"That's basically what I've meant. A possible solution would be:

less than 10k dps: no loot"



A shadowpriest on dispel duty, a rogue on interrupt duty, an equally skilled player with slighly less gear all those players would "conribute less" according to your rules.

A high DPS moron standing in the fire eating away the healers mana is fine?
Just looking at a single aspect of your playing won't suffice. There is so much to factor in, you can't make up a simple score.

Kring said...

> There is so much to factor in, you
> can't make up a simple score.

Sure. But that's also true for the stupid Gearscore add-on and it was used widely in WotLK.

My argument was that players invented Gearscore. But they still distribute loot without factoring in contribution.

To me that means that people are fine with not considering contribution when distributing loot. Otherwise such an add-on would exist (and would probably be as broken as Gearscore) but it would be used nevertheless. Such an add-on is not used and therefore it's stupid for a company to fix something that's not broken.