Greedy Goblin

Monday, January 26, 2015

The dumbest stat

CCP told on day 3 of the Winter Summit that "Kill per logon minute is increasing". I couldn't help saying "idiots" loud to the monitor. Why are they completely wrong?

Meet Yura Movsisyan. On 6:42, Jan 23 he decided to leave Dodixie. The local Deadly Fingertips camp disagreed, killing his worthless noobship and his empty pod. But Yura isn't a quitter. Until 7:36 he lost 15 noobships and pods, contributing 0.0003B to my latest GRR war. I doubt if the killers are particularly proud of their accomplishment, or that Yura will provide tears over the loss. This is nothing but cutter on the killboard and I wonder why there is no filter on zkillboard to hide irrelevant kills and losses.

Yet this "battle" had very high "kill per logon minute" and will affect CCP decisions. Probably Yura is placed to the "target audience" group. His activities are checked for correlations to figure out how he made the improvement from boring carebear into the PvP juggernaut he is.

Kill count is a completely useless stat. PvP must be measured by its consequence: ISK loss. All developer decisions that are made with the goal to promote PvP must be checked for its effect on ISK loss, not ship kills. Why? Because PvP means "player versus player". Two players fighting each other with the intention to win. Not a bored camper popping autopiloting noobships while waiting. Not drunken idiots killing time in frigs. Not careless noobs getting lost and popped in starter crap. If CCP forgets that and tries to optimize features for ship kills, they'll get angry birds in space. After all, that game has pretty high "kill per logon minute".


Pheredhel said...

If you say the stat is useless, you would implicily say that only noobship kills did increase. Where is the evidence that the mix of killvalues changed?
And even if it changed, where is the evidence that it was just the same players beeing killed more?

If you can't proove both, then it is not a useless statistic. It is not the information you were looking for. But it can be used as an indicator that either more people participate in PVP or that the ones who do are more active now.

For those values "more ISK destroyed" is a very bad statistic, as it can just be "people fly more expensive ships". That is a number that gives in world impact, but only when you have it as a ratio to isk earned. ISK destroyed could be inflation as well. I doubt you can find a single usefull number that tells it all.

The number seems to be a good indicator that activity in lowsec has increased.

Anonymous said...

*Isk is one metric,
*Strategic objective is another metric,
* Kills another metric,
* Participation "links" another metric.

You can measure the data how ever you like. But saying kills is not a good metric since ISK is better is not true. ISK counting would give the one killing a rare Tournament frigate full points while someone who killed 40x normal frigates almost no points.

In nullsec you count several metrics, Strategic objective is the more important one. What was our goal and did we meet it.
the Isk count is just to see did we meet the goal with flying colors or not.

Participation lists is good to have metrics on corps who did there part and who didn't It's also a good way to find the logistic pilots who sacrifice there own stats for the fleet.

Killstats in some kind is a good metrics for self stats. How am i doing compared to others.

Gevlon said...

Strategic objectives are good for an alliance, but not CCP, since for them, both sides are equally players.

CCP already measures participation: logged in time.

Killing 40 T1 frigs is less relevant than killing a single T3. The second was a fight, the first was a bunch of kids being silly.

dobablo said...

Most entertaining thing in EvE is bangs while online so I don't see it as being a terrible metric.
ISK destroyed is also meaningless because it doesn't cover the impact to individual players. Losing a T3 means nothing to a player with billions in the bank, while losing a frigate lost to a new player can be a set-back. My metric of choice would be percentage of assets destroyed per logon minute.

Anonymous said...

"Killing 40 T1 frigs is less relevant than killing a single T3. The second was a fight, the first was a bunch of kids being silly."

Or the second could be a guy AFK at a gate.

You have deemed T1 frigs to be useless and irrelevant. Your opinion does not necessarily match that of others.

Von Keigai said...

PvP must be measured by its consequence: ISK loss.

No. You have to look at what CCP is trying to do, to see what they should measure.

What is CCP trying to do with EVE? They are a business. They are trying to make money. Not ISK -- real world money.

How does CCP use EVE to make money? They do it by attracting and retaining players. So they want to measure stats that relate to that. So why do people come to the game and why do they keep resubbing? Because they are having fun. Yes, fun -- completely unmeasurable. But perhaps they can proxy. What's a better proxy for fun, explosions per hour or ISK lost per hour? Most players think it is the former, not the latter. Evidently CCP, who might be expected to know more about their game than you do, agree with the majority of players.

daniel said...

as the amount of isk in this game is increasing over time (i assume we do not have to argue about that) players become more risky at taking expensive ships into a fight, therefore the average value of killmails increases.
but this result absolutley has no significance when talking about general pvp activity.
when i started to play eve, i once lost a bestower, t1 fitted with no cargo. a year later i lost a prorator, t2 fitted with 2.5b cargo. at which time were i playing eve more activly?

unless you generated a (valid) statistic which proofs an unproportional increase pf "irrelevant" killmails, your thesis stands absolutely no chance at all.

maybe you tried to make a case about general activity in this game, in which you could be right if one assumes that it requires more general activity to farm one t3 compared to 100 frigs. but, this one t3 was farmed and flown by one player, while 100 frigs were farmed an flown by one hundred players, or one player one hundred times.

for comedy, let us assume, that both, the t3 and the 100 frigs are reasonably (t2) fitted which would put the t3 (no expert at all, just a rough guess) at, let's say 500m, while a t2 fitted frig would be, let's say 10m. 10m times 100 is 1b, therefore 100 frigs, in whichever way you put it, indicate towards more activity.

two things remain. one, it is getting more and more easy to make isk, and two, you missed to give a proper definition about what you were wanting to proof with your thesis.

Esteban said...

Yura is an outlier. And, as you well know, EVE PvP is rarely 'two players fighting each other'. Most kills, even between PvPers, will always be closer to a gank, because one will have the initiative.

Kill/logon minute (in addition to being good PR because it perpetuates the 'EVE is a jungle to prove yourself in' myth) is a decent supplementary stat describing participation specifically in PvP. It means that people are getting the little dopamine hits they set out to get when undocking in a PvP ship, and that's good for business.

I'm sure it's not the only stat CCP look at.

Gevlon said...

@Von Keigai: if so, then why do we have losses?

Wouldn't it be more fun if the destroyed ships would just respawn at the graveyard and could warp back to the fun?

daniel said...

Wouldn't it be more fun if the destroyed ships would just respawn at the graveyard and could warp back to the fun

actually there is a reason why many eve players do not like and want to play wow

daniel said...

ps, i just see, it's 40 t1 frigs, not hundred, my bad, my eyes ... :(

maxim said...

CCP are game designers. Their job is providing optimal experience to individual players, not optimizing ISK wars.

Rather, optimizing ISK wars can be a task within that job, but this is not the job itself. The job is providing optimal experience for individual players.

Kills per logon is a poor metric of ISK wars indeed. It is a decent metric of individual experience.

Gevlon said...

@Daniel: I'm fully aware. However what's the difference between ship respawn and permanent loss of a trivial cost ship? After all, in WoW you also have some trivial repair cost.

daniel said...

the diference is, that ccp is selling it's players the illusion that loss matters, while blizzard is selling the illusion that progress matters. ***

on a sidenote, when you diss us t1 friggers (yes, i'm one of them, though i suck horribly), you do not take into account that the loss of a t1 frig isn't just 350.000 isk plus insurance minus payout, but ~ 300k isk, plus aprox 10m in modules/ammo, which aren't recovarable (assuming you die (alot) and don't have the possibilities to loot your own wreck.
while 10m is nothing to you, for some players it is enough to convert the possibility of a loss into a tiny bit of additional adrenalin-rush while engaging a hostile pilot. ***^2
it would be really nice of you to acknowledge that you are not the only eve player who is logging into this game to spend a joyful time in a virtual space world. and much like you, other players as well chose to pursue their (more or less) unique path through new eden.

you also want to understand, and appreciate, that the frigate thing is causing a constant flow of destruction, leading to a constant need for supply of goods - which drives the thing that makes eve so unique and in the end enables you to follow your individual playstyle ... a really huge player driven economy/market.

*** in reality they don't matter at all, because videogames.

***^2 in reality 10m should be nothing to allmost noone ***^3, and i think ccp designed the game like that on purpose. think of t1-pvp-pilots a bit like, e.g. unreal tournament players, but in a bigger and more complex arena with engagement mechanics that cater less for the fast-aiming but more strtegic thinking player. the frigate game is a minigame within the game. like isk grinding (and supporting moa) while rampaging against specific alliances is your minigame within the game.

***^3 10m aren't really that much, but 5 or ten times 10m (a day) are significant - for players who prefer to use their (limited) gaming time for something other than grinding isk.

Von Keigai said...

Gevlon, we have losses it's true, but the situation is not a boolean one of kill/killed, win big/lose big. Rather, the player controls how much his loss matters to him, by what he flies. If you want a game with lots of explosions and lolling about, you fly frigs. This is not exactly "the destroyed ships... respawn at the graveyard and could warp back to the fun"... but close. A frigate requires only a few minutes of work to pay for. If you want a game where losses matter, you can fly a T3. Then when you die you lose not only the 500m but 4 days of training. Woohoo. The point here is that EVE is large enough to encompass both playing styles. There's ISK risk for those that find that fun. And there's explosions and almost free respawning for those that find that fun. Neither play styles is right or wrong. Some people like one, some the other.