PDA

View Full Version : Beyond Kill Score



Dunfalach
08-23-2010, 10:03 AM
I understand why we have a kill score metric on the XP report, and it's a useful metric as far as it goes, but I think two other metrics need to be added to go with it:

1) Kill assists. Only one person can get credit for the final kill, but often many people may contribute to the damage on any one enemy unless it's a one-shot-kill situation. If someone shows on neither kill score nor kill assist list, they still might have contributed highly if they're the buffbot, healbot, etc. But it spreads the credit out more accurately on the kills.

2) Damage dealt. It is entirely possible, in theory, for a person to do massive amounts of damage and yet have few kills, especially if they're a caster doing area of effect spells. Addiding damage dealt to the report (even moreso than kill assists) would help in determining whether a low kill score means your DPS is lagging or simply represents others getting the finishing blow in. The combination of kill score and damage dealt would make the metrics much more useful in evaluating performance.

Now that I think about it, it wouldn't be bad if there were a way to add a hp healed metric for clerics and other healers.

I figure all the above are numbers the game is already tracking, so hopefully not too difficult to implement. They may be more complex to achieve than I realize, though.

Summon and charm players will obviously have to rely somewhat on the "Misadventure" scores to get an idea of their numbers. And the contribution of buffing and immobilizing spells would not be easy to metric at all, I would expect. But those are more easily seen by everyone. Knowing who's causing what damage is much less obvious when you may have 6 people all attacking the same target and can only see your own numbers flying by.

I think adding at least the damage dealt metric would improve the utility of the report. The others would be nice to have too, but if I picked just one it'd be damage dealt.

ghettoGenius
08-23-2010, 10:09 AM
This has been suggested probably about 1000 times. It has also been rejected just as often as it is generally believed to require too many resources to store and calculate running DPS totals.

k1ngp1n
08-23-2010, 10:14 AM
So, I studied statistics quite a lot in my math minor in college. :)

Kill score correlates quite nicely with damage done. Over time it will trend to match. Yes, kill snipers can cause a confounding, and will sway the numbers. But in a generic quest in which everyone attacks normally, kill scores = percentage of total dps.

There is absolutely no need for these metrics.

/not signed.

Hjarki
08-23-2010, 10:14 AM
I understand why we have a kill score metric on the XP report, and it's a useful metric as far as it goes, but I think two other metrics need to be added to go with it:

The actual reason we have a kill score is because it's necessary for the game to track kills in order to award bonus experience. It is not, as you seem to believe, as a metric for players to measure their performance.

As far as I know, no major game performs metrics of this sort for players - all of the 'dps meters' are third party applications of some sort. So really what you should be asking for is the ability to save logs (which is how games other than WoW, which has an extensive macro language, do it).

Cam_Neely
08-23-2010, 10:15 AM
Well, the Devs have a limited amount of man hours to work on the game. I would rather they put it towards bugs and content as opposed to E-peen markers.

Anyone with half a brain knows that the kill count does not matter. Quite often on my Barb (who i just Capped over the weekend :) FIRST one!) I have 50%+ of the kills. But I know as well as anyone that without the healing aura or heals, I would not be running frenzy and death frenzy. Without the thf running around and paralyzing 4 mobs to line up for me to take down, the quest would go slower. Without me the rogue would do less DPS. With out the caster I would not have haste, web and firewall helping with the kills.

And so on.

KillEveryone
08-23-2010, 10:39 AM
So, I studied statistics quite a lot in my math minor in college. :)

Kill score correlates quite nicely with damage done. Over time it will trend to match. Yes, kill snipers can cause a confounding, and will sway the numbers. But in a generic quest in which everyone attacks normally, kill scores = percentage of total dps.

There is absolutely no need for these metrics.

/not signed.

Your kill score can be assisted greatly by symbol of stunning, symbol allowing for auto crit, as well as someone that is using a weighted item with stunning blow which will also allow auto crit allowing someone else to get the kill.

There are other thing happening in the group and the group dynamics make things a bit more complicated.

Kill count doesn't truely show the contribution of a battle.

That being said, if it is easy to actually show battle stats that give a good picture of what happened than fine by me but if it is going to be too complicated to implement then I'd rather the Dev's work on other things such as...

bug fixes, content, pre's, and all those other goodies.

Dunfalach
08-23-2010, 12:13 PM
The actual reason we have a kill score is because it's necessary for the game to track kills in order to award bonus experience. It is not, as you seem to believe, as a metric for players to measure their performance.

As far as I know, no major game performs metrics of this sort for players - all of the 'dps meters' are third party applications of some sort. So really what you should be asking for is the ability to save logs (which is how games other than WoW, which has an extensive macro language, do it).

When you say bonus XP for killscore, you're just talking about the bonus XP everyone gets for wiping out enough monsters in the quest, not an individual bonus for number of kills, right?

Possibly no major MMO does. My MMO experience is limited to DDO and the LOTR beta. So I'll defer to your greater experience on that. It's a very common metric in FPS games. The Battlefield games, for instance, track kills, kill assists, healing someone, repairing equipment, handing out ammo, destroying objectives, capturing flags (in a CTF mode game), hit percentages, and other factors. I'm not suggesting we need that extensive a list, but I'd love to see at least damage totals. The last hit is always the hit that kills, but not always the one that did the most damage.

AyumiAmakusa
08-23-2010, 12:20 PM
This has been suggested probably about 1000 times. It has also been rejected just as often as it is generally believed to require too many resources to store and calculate running DPS totals.

Yes.

Also, who in the world cares about the kill count?

And no, clerics don't care that they healed the entire group for 5000 HP or even 50000 HP.

Dunfalach
08-23-2010, 12:21 PM
This has been suggested probably about 1000 times. It has also been rejected just as often as it is generally believed to require too many resources to store and calculate running DPS totals.

If it requires too many resources, I can accept that as a valid reason. Although, there may be a separation between calculating damage per second, vs just keeping a running total of damage. Measuring damage against seconds would seem to be more resource intensive, since it's a very rapid running calculation, then just tracking the total.

kernal42
08-23-2010, 12:26 PM
So, I studied statistics quite a lot in my math minor in college. :)

Kill score correlates quite nicely with damage done. Over time it will trend to match. Yes, kill snipers can cause a confounding, and will sway the numbers. But in a generic quest in which everyone attacks normally, kill scores = percentage of total dps.

There is absolutely no need for these metrics.

/not signed.

This is only true in a perfect world.
And DDO, we all know, is not perfect ;).

Counterexamples:
Wall of Fire expires. Caster did tons of damage, gets no kills.
Fighter enters melee activating haste boost; does lots of damage, but haste boost expires before mobs are dead. During the "killing blow" time, his dps is far less (resulting in lower chance of kill) than would reflect his damage dealt.

On the other hand, counters like "damage done" are very cheap in terms of system resources.

Personally, I'd prefer a tongue-in-cheek total counter.

Helpfullocity: Damage dealt + saves made + traps disabled + damaged absorbed (DR) + distance jumped + enemy saves failed + kills + damage healed + debuffs removed + buffs cast + sp used + collectibles taken - falling damage taken + .......

Cheers,
Kernal

k1ngp1n
08-23-2010, 12:48 PM
This is only true in a perfect world.
And DDO, we all know, is not perfect ;).

Counterexamples:
Wall of Fire expires. Caster did tons of damage, gets no kills.
Fighter enters melee activating haste boost; does lots of damage, but haste boost expires before mobs are dead. During the "killing blow" time, his dps is far less (resulting in lower chance of kill) than would reflect his damage dealt.

On the other hand, counters like "damage done" are very cheap in terms of system resources.

Personally, I'd prefer a tongue-in-cheek total counter.

Helpfullocity: Damage dealt + saves made + traps disabled + damaged absorbed (DR) + distance jumped + enemy saves failed + kills + damage healed + debuffs removed + buffs cast + sp used + collectibles taken - falling damage taken + .......

Cheers,
Kernal

Over time, these all balance out.


Over time it will trend to match. Yes, kill snipers can cause a confounding, and will sway the numbers.

*Sigh*

WoF is just as likely to kill as anything else. Only time the kill score doesn't match damage done will be when melees wait to start dpsing until some time after the mob entered the firewall.

Haste boost can go both ways, and lasts a lot longer than it takes to beat down a trash mob. During that time their kill rate will be higher.

Stunner builds will still have their kill score match their dps. Contribution of the stun is a separate issue.


Kill count doesn't truely show the contribution of a battle.

I said dps and kills, not overall party contribution.

This game doesn't need these metrics at all. We've done fine without them, no reason to waste dev time on useless counters to either show of e-peen or to encourage more of the kill-score centric play style still so common amongst new players.

still /not signed

Cylinwolf
08-23-2010, 12:52 PM
And no, clerics don't care that they healed the entire group for 5000 HP or even 50000 HP.

Personally I would be laughing my ass off into the mic if I managed to get a 50,000 heal crit.

kernal42
08-23-2010, 01:01 PM
Haste boost can go both ways, and lasts a lot longer than it takes to beat down a trash mob. During that time their kill rate will be higher.


Haste boost can go both ways, but I imagine not many activate haste boost when the trash mobs are almost dead. More likely it's activated near the start of a fight. It either lasts the whole fight (high damage & high kill count), or it ends early (high damage & lower kill count).

Of course, I'm really only nitpicking. I certainly agree that the correlation is strong enough that kill count ratios approximate damage ratios. Boss fights also skew this, and damage vs. boss would be the most interesting counter, but also the biggest waste of time for Turbine to implement.

Cheers,
Kernal