Greedy Goblin

Wednesday, April 2, 2014

The inverse effort/skill vs reward problem

This post and the following discussion helped me a lot to understand the "PvP-er" mentality, and also to understand a serious problem with the EVE learning curve, resulting the infamy of the game.

In the post Von Keigai told us his long and finally successful hunt for an Iteron V in wormhole space. I commented "I am completely at loss why did you waste so much time trying to gank an Iteron. If your heart desires Iteron killmails, you can get them by dozens in Uedama." he replied "my heart does not desire Iteron killmails. It desires killmails where my gankee was trying hard to avoid me, and yet I managed to hunt him down anyway. See, ganking someone who's defenseless and not trying to avoid you is basically an engineering problem. It's solved. It may be an interesting challenge to solve that problem, and it may be interesting to optimize the solution. But once solved and optimized, it's no longer an interesting process to run; it's just grinding. You should know. Obviously ganking highsec miners is not interesting or you'd still be doing it."

This is a very good explanation of "try real fights instead of ganks". I used to dismiss that argument as lie, since vast majority of kills in EVE are ganks where the target has no chance to win. I understand now: while that Iteron had no guns, it still could and did fight back by cloaking. "Fighting back" is defined as "using intelligence, skill and effort to resist the attempt of the attacker".

So, the goal of the "PvP-er" is to face an opponent that has both skill and effort to resist the attack and defeat him using more skill or effort. Ganking a miner in Wormhole space is a PvP victory because you have to find him without him being alerted to your presence. Ganking a same mining ship in highsec is just a gank because he was AFK in a belt where everyone could see him.

There is a fundamental problem with it. No, not my usual counter-argument that "it can't be proved". While it indeed can't be proved, the PvP-er doesn't want to prove it. He plays for the moment of victory and can't care less if I kill 10x more identical ships after I paid their owners to die. The problem is that this creates an inverse effort/skill vs reward curve. What does it mean?

In every single game your results improve with your skill and effort. If you play League of Legends better, you get higher rating. If you play World of Tanks more, you get more XP and credits. If you are a better raider in WoW, you get more epics. "Dumbing down" or "casual friendly" decreases the difference between good and bad players, helps scrubs to continually catch up, but the better one always have more game rewards. On the other hand in EVE playing better can and often does give worse results. Why? Because obviously bad players are ignored by PvP-ers, while not so bad ones are attacked. That Iteron would be still be alive if it was autopiloting in Uedama, because its cargo didn't warrant an economic suicide gank. He died for only one reason: he was better player than those who autopilot in highsec and went to live in WH space.

There are two results of this behavior: the first is frustration of new (but not very new) players. They learn and learn and get worse results. They don't understand what happens, why aren't they improving despite their efforts. The truth is that they are improving, but they are facing much stronger opponents. It's like a rating system, without rating. It's obvious that winning a Platinum 1 match in League of Legends is harder than winning a Bronze 5. However it's expected and the progress was already rewarded by the Platinum badge. The progressing EVE player doesn't have a badge or rating to know that he is progressing and just lost a "Platinum" fight, which is better than winning "Bronze" fights. The game doesn't give any clue that he at least attempted something hard. He cannot know - even less prove to others - that he is any better than a 1-day old newbie. I remember how frustrating World of Tanks was until I figured out that the random matches aren't random and a hidden rating system places me in a position of 50% win rate regardless of skill.

The other outcome is that players who don't aspire for peer respect, but objective results can get these results unopposed and in stellar amounts. If you optimize simple stuff, you can get awful lot of ISK or awful lot of kills. This is true even in the most extreme form: pilots who are so dumb that they aren't even human (AFK-ers, bots) can be among the most profitable pilots. For such players EVE is a trivially easy game. I was PLEX-ing my account on the second month and got 10B by the third and look puzzled why everyone is calling EVE hard. When I got enough of "try PvP" comments, I went and got more kills alone than medium-sized alliances. Anyone with my attitude can easily come to the conclusion that most EVE players are awfully bad, even by WoW standards. I mean that a "horrible" WoW player has 1/3 damage of a good one and not 1/100. Actually they aren't bad, they just play a very different game, where the goal is defeating another, challenging opponent.

Too bad that the game system does not support this kind of play by a visible metric. So I must ask: why do they play EVE? I mean one could try to become the "best duellist in WoW, killing others in duels front of Orgrimmar", but without any in-game support, that neither can be verified, nor has a wide scope (many good opponents never go there since duelling is stupid). There are lot of games where personal skill is directly translated to ratings and prizes. Why don't they play Chess or Starcraft instead?


Penny Nickals said...

Eureka!!!!! Love your metric driven approach most things, always thought your approach has always neglected the above. Well said.

I'd challenge: Is "While it indeed can't be proved" really true? Really seems like a data problem, with enough of the right data, there should be a way to assess quality of kills. And with such a metric, you could right the inverse effort vs. reward curve.

Anonymous said...

Interestingly enough, EVE is probably your biggest challenge yet (in that it is such a social mentality driven game).

You will continually run into the "sandbox" vs. objective metric problem.

The fun part will be in how much of an impact you might have. I'm not talking vs. Goons or in the debates (lack of?) with your commentators, but in verifiable changes to the "line member" or average joe of EVE, and how he/she begins to make rational decisions towards whatever goal chosen.

The biggest problem? EVE is still more akin to a chat room with a sci fi/exploding spaceships skin. For better or worse.

Anonymous said...

that moment when the slowest horse finally crosses the finish line...

glad to hear you've finally worked out why killing your own carrier with an alt didn't earn you any respect

Foo said...

Why don't they play Chess or Starcraft instead?

Chess, Starcraft have an implicit reset to zero every game. While your game history affects the opponents you have access to, or what 'soft skills' you learn, you start chess in the same position, either white or black.

Eve instead is a persistent universe where what I did yesterday affects what I do today.

Using a sandcastle analogy:

In chess I start every game with a fresh pile of sand.

In Eve, I spend spend time improving my existing sandcastles, and building additional ones.

Whether I play chess or Eve, there is someone wanting to knock my sandcastle down.

In chess, win or lose last round, I start with a new pile of sand.

In Eve, if I won last round, I get to build on what I already have.

Anonymous said...

"There are lot of games where personal skill is directly translated to ratings and prizes. Why don't they play Chess or Starcraft instead?"

Here's the answer: "Because they find it fun."

The only question now is, since your personal philosophy can't seem to accommodate "fun" without apparent tangible reward, is whether your head will explode or whether you'll just dismiss the point out of hand.

lasse said...

FINAlly you geht it :-) Eve is AS hard AS you make it ... You are your own rating system wich makea Eve one of the most Boeing and most entertaining games aroud Luke kerbal space system ... You get your joy from getting to your goals

Arrendis said...

It’s another 2-parter, Gevlon… sorry

If you play World of Tanks more, you get more XP and credits.

And at the same time, please note that at higher tiers, you need significantly more xp and credits to make incremental improvements in your tank. No matter how skilled you are, it's not impossible to fully upgrade your Tier 1 leichtraktor after 1-2 matches. Try doing the same in a Tier 9 M103.

Proportionately, you are making far, far less in terms of rewards as you approach the higher end of the invested time/effort scale. This is hidden by the use of numbers that make the rewards appear larger, but every player knows, ultimately, that the more you improve, the more your continued improvement on any given amount of gameplay is reduced. It's a set of diminishing returns, just like most actual systems tend to be in the real world, as well.

Because obviously bad players are ignored by PvP-ers, while not so bad ones are attacked. That Iteron would be still be alive if it was autopiloting in Uedama, because its cargo didn't warrant an economic suicide gank.

I think it would be better to say that hunter would not have killed that Iteron if he was autopiloting in Uedama. You are still more likely to be killed autopiloting in Uedama at any given moment than you are if you're flying smart in a given W-space system at that same moment.

If you optimize simple stuff, you can get awful lot of ISK or awful lot of kills.

And now you start to see why we say that killboards are meaningless, why the numbers you've been touting for the 'Grr Goons' project represent nothing as far as the CFC's concerned - those people who got caught should have been caught. Their losses are nothing in the grand scheme of things, no matter how impressive the isk numbers look on a killboard.

This is why BNI's willingness to throw themselves headlong into the jaws of death gets accolades from the big null blocs, when Lemmings get mocked for their insistence on running away from fleets where they could have given a good account of themselves: because the 'isk war', on these low, low scales - basically anything under a few hundred billion - is meaningless. What's more important really is the 'content' - are you stepping up to fight? Are you fighting smart? Are you punching above your weight, or below it? Are you continually testing yourself, again and again...

... and if not, why are you here? Why are you playing the game? Aren't you bored? Find something interesting to do. Challenge yourself. For the love of god, don't just be the EVE equivalent of a couch potato.

Arrendis said...


Too bad that the game system does not support this kind of play by a visible metric. So I must ask: why do they play EVE?

Because game-supported ratings mean as little as killboards. Think about it: You've just identified that the game-presented numbers of 'this is worth X' aren't what drive PVPers. Beating people starts to get there - beating a challenging opponent at the game, not putting up meaningless numbers that can be racked up just as easily by, let's face it, utter morons.

I don't run fleet logistics so I can keep people in formation when we're sitting around babysitting a capital fleet. I do it because when I need to keep the logi in the right position, weaving around multiple enemy fleets on a broken grid, horribly outnumbered and fighting tooth and nail to keep as many of my fleet alive as I can... that's fun. That's hard. It's challenging, and I have to stay sharp. How would you even begin to establish a game-based rating for that?

And there's the difficulty: How do you even begin to estabish a metric for how well you perform in situations that don't even approach an even fight? In order to give a rating on someone's performance, you have to have a baseline to measure it against.

Let's look at your example, and that Iteron. How challenging is hunting that Iteron? How do you even begin to measure his performance? How long it takes him to die? So the 'best' ratings go to... people who cloak up and don't get killed, so they never show up in the first place?

I have my own metrics for my performance. Most PvPers do. My own metrics include the performance of my logi, sure, but they're also highly situational - different situations mean there's a different level of success I can hope for. If I'm outnumbered 50:1, I'm probably going to have a much worse day than I'd like, no matter how well I do my job. And it's impossible for any mechanical rating system to take all of that into account.

The reason games like WoW and Chess and Starcraft can establish competitive ratings is because they begin from the assumption that the game board starts off even. Neither side has an advantage before the game begins.

Eve laughs at that assumption. Eve mocks it, openly, for the preposterous notion that it must be in any persistent, continuous game. Games of Chess begin, and end. Games of Starcraft begin, and end. PvP in WoW is a mini-game that begins, and ends.

PvP in EVE began. Eventually, it will end. So far, the first match played is going on 11 years. So maybe, maybe in a hundred years, we'll have six or seven matches done, and we can begin to establish rankings for... what is it, half a million or so? Three quarters of a million people?

I don't think seven matches is really going to give us a decent season. We'll have to play the full 162 games.

But hey, that means around the year 4350 A.D. or so, we should be able to establish those ratings and prizes you want.

Anonymous said...

A few notes:
Bad players are not ignored, there are bad pvp-players who boost their stats using 'Em, and there are pvp players who see it as a duty to kill those who don't learn. Basically not all PvP players are alike.

Then the progression thing: it never felt like that to me. I started PvP only being able to go after very specific targets. Over the I learned what I need to know to be able to successfully and safely pick fights with nearly all sorts of ships, so clearly I improved.

Arrendis said...

and there are pvp players who see it as a duty to kill those who don't learn.

You know... like you did, Gevlon.

Smith said...

You go from not understanding to hearing the talking points to trying to quantify them.

Individuals feelings do not have metrics. It might appear that way, but that is only correlation.

When you try to define a metric like "become the best xyz" you are falling into the trap that you did when you started the game. There can be no metrics for 'fun'.

I hunt extensively in intervals. I will attack anything that I have a remote chance of killing. Like vK I'd say that it's the thrill of the chase that is the main thing, but even though a chase not ending in a kill can be fun, one ending with a kill is more rewarding.

I don't get how you can reason that 'worse' players would get away more as they would prove no challenge. That is not what happens. Less experienced players will just be caught faster. Few hunters I know spare targets. Personally I kill them all and the really new ones I reimburse and try to tip what they did wrong.

Side note: why is it that you feel the need to measure everything? Is it a 'what can be measured can be proven and therefore real' thing?

Gevlon said...

@Penny Nickals: it's more than data gathering problem, it's rather existence problem. You can only qualify a kill if you qualify the opponent. So a rating system is needed, which automatically removes the right of the killer to "find fights" as in any rating system you must fight the proper enemies.

@Foo: interesting point, the persistent element.

@Arrendis: but the value of the reward is increased. I mean having a single T10 is better than having all the T1-2 tanks in WoT.

No, it's factually not true that you are more likely to die in Uedama. You saw the destruction graph, highsec is a little green dot in the middle, despite majority of people live there.

You are wrong about the ISK war. There is a reason why CFC used bombless bombers instead of titans. There is a reason why B-R is not repeated. Finally, there is a reason why a PLEX is only 700M, making a bombless bomber costing less than a cup of coffee: people are poor.

I'm here because I have nowhere else to go. All other MMOs openly cater to the lowest common denominator.

Proper rated games (like League of Legends) both provides you people to defeat and challenging opponents.

"Your own metrics" = "I'm awesome because I say so"

@Smith: worse players get away by not doing anything theoretically huntable (ratting bot in nullsec, warping when someone enters local) or they stay in highsec where they are safe unless they haul too much. They are safe because killing them is "just ganking"

Smith said...

So what is 'worse' about not getting killed? I'd say that if something is good or bad is measured in success in what you set out to do, agreed?

So someone hiding whenever someone enters local need not be bad/inexperienced or however you define it. That player might just be goal driven and figure that when mining in Null hiding is a smart thing when someone enters system.

I seem to remember you advising POSsing up and/or logging out when hunters where in your wormhole. DOes that make you bad? risk averse? Smart? Or what?

Arrendis said...

Here we go again... P1:

You are wrong about the ISK war. There is a reason why CFC used bombless bombers instead of titans. There is a reason why B-R is not repeated. Finally, there is a reason why a PLEX is only 700M, making a bombless bomber costing less than a cup of coffee: people are poor.

I am not wrong. There is a reason the CFC uses siege bombers: they're easy to insert, easy to extract, and difficult to catch, while still doing reasonable DPS to large, immobile objects. They're also far easier to field in large numbers than titans are, meaning there's less of an organizational hassle in getting the requisite numbers of pilots engaged.

And I know you really, really like the 'bombless bomber' phrase, but you know it's factually wrong, right? Siege bombers have their bomb launcher offlined - it's useless against a POS, and there's no point having it powered up to fat-finger and launch, but it's there, and it's loaded.

As for why B-R isn't repeated... did I not say basically anything under a few hundred billion? I'm pretty sure I did. How many trillions were lost in B-R? Never thought I'd have to say this, Gevlon, but trillions > billions.

Siegefleet isn't about people being 'poor'. It's about using a tool suited to the job you need done. Siegefleet can bridge into a system, jump the BlOps in behind them, hit a tower, and jump right back out. If anyone comes into the system, the BlOps has already been cloaked up in a safe spot, and everyone else can just scatter and cloak, allowing the fleet to be difficult to catch, and frustrating to try to chase.

Continuing to harp on siegefleet the way you do simply demonstrates a terrible understanding of the dynamics of the situation. Tell me, when you need to open up a can, do you blow it up, or just use a can opener? There is such a thing as over-engineering your solutions, you know.

As for PLEX prices... PLEX prices fluctuate on a fairly regular cycle. When the demand is high, the price is high. If the demand were higher, the price would go higher. There's no shortage of supply, after all. So the pricing is entirely a matter of demand.

having a single T10 is better than having all the T1-2 tanks in WoT.

Better in what way? The T10 is a money loser, not a money-maker. It's limited in what engagements it can participate in. Its ammunition is expensive, and the matchmaker will continue to make sure your random matches are pretty much a lottery draw in terms of your chances of winning any given match.

Is it a more powerful tank? Sure. However, relative to the opponants it will face, it's not appreciably more powerful than say, a Tier 6 M4AE8 is, relative to the opponants it will face. And for this, much like with capital ships, you have a less kinetic gameplay experience with less action.

Personally, I prefer my VK3601 to pretty much any of my T8+ tanks. It can play as power, or it can play as active.

You saw the destruction graph, highsec is a little green dot in the middle, despite majority of people live there.

You really do fit the data to your conclusions, don't you?

Highsec sees less value destroyed, yes. Tell me, Gevlon, how many times in the last six months have multiple max-sized battleship fleets fought in highsec?

How many carriers have been destroyed in highsec this month?

How many supers have died in highsec this month?

The most expensive ships in the game don't go into highsec. The valuable fleets don't bother fighting in highsec. This skews the numbers, and there's no way around it. To claim that because the numbers say most of the isk destroyed in the game isn't being destroyed in highsec means that the kill distribution curve is the same across all ship classes is just nonsense.

Anonymous said...

@Gevlon, but why don't you play Starcraft or any other competitive game? It's not even like you can only win 50% of the time, as at some point there won't be enough players around who can consistently beat you.

jstk said...

What you see is another symptom of one person always "winning" eve depending on ther point of view. The pvp'er is winning on his mind because he found and killed his target. If his targets runs away, the hunter think he "won" because he made the other guy go away, while the sheep also "won" because he didn't give away an easy kill. Since that guy didn't even link back the Iteron killmail I assume it was empty so the victim probably won on his mind as well as he "didn't want that Iteron anyway".

Also you are incorrect with your world of tanks analogy. A 50% matchmaker does not make everyone look like a winner, it's made so everyone can have a relatively fair game with people around their level. A matchmaker witouth an internal MMR system is useless. A player with 50% win rate at 5000 MMR is, by definition, better than a player with 50% win rate at 2000 MMR. Your individual effort does matter in the long run.

As for Eve, what you describe is a symptom of the "sandbox" issue. While it's nice to have freedom on what you do, the only acknowledgement you get from it comes from what you convince yourself and others to think about it. It's impossible to objectify what's a good kill or a bad kill when there are no equal circunstances for both teams and because the game plays on a persistent world (where the playing field is always evolving).

Why do you think propaganda is such a powerful tool in EvE? Because the only winning condition is on your mind.

Anonymous said...

The most interesting post since a very long time... To answer you last question I think that this is precisely the abscence of metric that appeals to EVE players.
Every pvp pilot is persuaded to be above average because he can choose the metric he wants or include other parameters "they had us outnumbered" ; "we fought for fun this time". Starcraft, chess of LoL are harsh games, they tell you without a doubt that you are bad, in EVE nothing tells you that and often enough there is even a community to proclaim how awesome you are (see goons, N3, low sec, wh community).

lacal said...

I'm pretty poor at competitive games, with EVE I can tip the scales heavily in my favour with forward planning, so with less skill I can get the reward.

The reward for me was always the 'solo' big isk killmails and associated loot drops, gained through said forward planning, it made me feel 'smart' - upwards of a year ago this really didn't require much effort either.

Eve today requires more effort for less of my chosen reward due to the bulk of lo-sec being low isk value targets, sure I could go roaming and hunting for the T3's etc. but thats the increased effort.

So for now I log in once a day for 15 minutes and station trade to plex my accounts, once the effort required decreases or my chosen reward increases I'm sure I'll go back to PVP.

Side note - discovering station trading has made all other PVE content reduntant!

Anonymous said...

It might be the first reasonable post I've read here. So - I'll answer this.

The entire satisfaction of victory comes from the possibility of failure. There's immense satisfaction in catching someone who tried to evade you, in executing a plan that culminates in a strike that inflicts actual damage.

"Actual damage" is not a measure in ISK - it's how what you've done affected a corporation, or alliance. Has it caused internal strife that, in the end, ripped the corporation apart? Have your actions contributed to them abandoning an objective and moving out?

In the age of people owning billions, or trillions of ISK saying "I killed 100bn worth of goons" is nothing.

Possibility of failure comes from many factors - not flyin an iWin ship of the month in solo. Trying to catch someone who succeeded in avoiding getting caught, instead of picking at the lowest hanging fruit.

Anonymous said...

Good post, however what you are not trying to do, when you say that stuff is unprovable, is probably to measure the human mind.

Reminds me of a fellow in another imaginary universe, Hari Seldon, that created psychohistory, measuring the psychology of mankind and predicting the future.

Man is not as stupid as you say. They are just acting dumb because of many factors.

Anonymous said...

My favorite fight in eve, my fleet lost. By any metric of killboard efficiency, isk lost, field not held, we flat out lost.

Yet never was my heart racing more. Never was the thrill of the hunt greater, never did I face such a challenge. It was fun, and even losing thr battle, I would do it again. Granted, winning would have been better, but many fights I've won were not as fun.

As for the fight, my fleet was on the far side of a gate in nul. I, as well as our targets, was on the close side. I was cloaked in a bomber,trying to sneak to within scram range of the lone enemy blackbird. The plan was for me to stop cloak and attack thr blackbird, either destroying it or driving it off as a distraction as my fleet jumped in. I did drive the blackbird off and escape, but our fleet still did not win.

Have I ever felt that rush in chess? No. In chess it is me verse one other. There is no team dependant on me to run my portion of the plan. The same is true of starcraft. I enjoy FPS games, but none have had that rush either. The closest was in an old game Chromehounds, when sneaking around to kill the enemy base.

So you ask why I play eve, why I enjoy PvP even if I lose, and I say because no other game has the rush and the challenge on so many levels as eve. It's not based on isk made, it's not based on wins, kills,or the like. It's because of all the challenges and my enjoyment out of surviving and trying to beat those challenges. And yes, the challenges are set by my own standards, but that is good to. If I am feeling lazy, my challenge can be simple level 4s. For something harder, running the sisters of eve arc when wardecced by Lemmings and Marnite, all the way upto sneaking into a major nul fight to drop bombs and cause chaos.

Anonymous said...

You are wrong about the ISK war. There is a reason why CFC used bombless bombers instead of titans.

Yes, there is a reason. It has exactly nothing to do with worrying about massive killboard losses. It has everything to do with being mobile and frustrating the enemy. It also allows goon's to bring one of their greatest assets to bear - numeric superiority. An alliance which is built on adorable newbies can field huge bomber fleets far more easily than it can field huge titan fleets.

The proof, as they say, is in the pudding. Siegefleet objectively works. If it was as lolzy as you think it is it wouldn't work.

There is a reason why B-R is not repeated.

This is where you get into a different scale. At the scale of B-R you are right - isk DOES matter. Not so much because trillions of isk was lost, but because the real human cost in time to rebuild such a fleet (through the whole production chain, starting at raw materials and the time to extract them or haul them in) is enormous. Lose a siegefleet, replace it instantly. Hell, goon logistics have thousands of replacement ships for most of their doctrines moved into staging systems for wars. So at the scale of of 10s to 100s of billions of isk, losses literally do not matter. They don't even hurt.

Finally, there is a reason why a PLEX is only 700M, making a bombless bomber costing less than a cup of coffee: people are poor.

Plex rises and falls. 700M isk is a very high spike, but I think you'll find the "true cost" of a plex is being distorted by at least some amount of speculation. In fact basic supply and demand would suggest that it isn't because people are in-game-poor and injecting plex into the system (this would drive the price down)... it suggests that people are buying up plex that already exist in game and the supply is dwindling. (I've not looked at the numbers so I'll openly admit this conclusion might not be accurate).

Anonymous said...

I only skimmed the other responses, but on that skimming it looked like the same general idea. NOW you get it. Finally. Not from the Scrubs v Goons perspective, or the KB Efficiency perspective, or the ISK/hour ratio perspective, or from any other "objective" perspective.

I once spent about 2 hours lining up a WH kill in my cloaky Proteus. The ship I killed was in every respect inferior to my own, and once I got him tackled it was an easy gank. But the guy piloting it was so savvy and careful that it was a big achievement. He knew he could have been being hunted, even if he didn't know he "was" being hunted, if you know what I mean. I only probably made 10 mil ISK from the job, but it was so satisfying to finally outwit his precautions that it made it all worthwhile. THAT is EvE.