Some statistics on class balance

[Foolou]

After reading some wines about imbalances and overpowered classes I decided to do some statistics on my own. I wrote a little python module that polls the statistics of the Servers Horus, Ra, Muspell and Niflheim by realm and class (2160 chars in total) and did a little survey.

I calculated the rp differences of yesterday (sept 8th) to today (sept 9th), threw away all players that did not gain rp last night and did some simple statistics on the data, here are the results.

There were 44 extraordinary players in the toplists (rp-gains bigger than average+2*standart deviation), the classes of these are:

Barbarian:6
Knight:1
Hunter:15
Marksman:11
Conjurer:5
Warlocks:6

So the enthusiast players clearly prefer the bow classes, there is just one knight playing on server Muspell (Doggy-San), maybe we should ask him how knights need to be played to be effective.

But lets look at the broader mass of players (rp smaller than average+2*standart deviation):

Overall average rp: 151 (730 players)

(Class) (Average) ( played by ) ( total rp gained by class )
Barbarian 145 117 17038
Knight 112 101 11341
Hunter 180 148 26686
Marksman 181 128 23174
Conjurer 131 126 16581
Warlock 143 110 15777

I think overall it is safe to say that the bowclasses are a little stronger, knights a little weaker in the big player base, but the difference is not dramatic (!).

Somehow hunters and marksman become a very deadly tool in the hand of the very ambitious players, which seems difficult to achieve for other classes, especially for knights, but that's just guessing. There might be a other reasons.

The general problem with this survey is that there are a lot of inactive players in those top 30 rp charts (2/3 of the players listed did not gain rp during the last 24h).

Please don't start discussing about reasons for the results, these statistics do not tell anything about that, they are just numbers.

[xomad]

I must say that if you get results by subclass as you did, you see a lot of diference between ranged and non-ranged players.

That diference is more when you see results by class...

Warriors (Barbarian:6, Knight:1): 7
Archers (Hunter:15, Marksman:11): 26
Mages (Conjurer:5, Warlocks:6): 11

Overall...

(Class) (Average) ( played by ) ( total rp gained by class )
Warrior (Barbarian, Knight): 257 218 28379
Archer (Hunter, Marksman): 361 276 49860
Mage (Conjurer, Warlock): 274 236 32358

By your results...warriors are the weakest class.

Regards

P.S. - Nice work with statistics ;)

[ArchmagusArcana]

Consider also that most archers dont use areas, and are the only class that lack a mass buff or auras, all other classes have these, thus, far far less 'leeching' is possible. I know on barb OFTH or ONS will net a large gain of 1rps at a fort, and thats just one single buff, not even including areas. Just some additional food for thought.

[Eaten]

Consider also that most archers dont use areas, and are the only class that lack a mass buff or auras, all other classes have these, thus, far far less 'leeching' is possible. I know on barb OFTH or ONS will net a large gain of 1rps at a fort, and thats just one single buff, not even including areas. Just some additional food for thought.

This may be so, but archers in general are doing a bulk of the fighting meaning more RP per kill and more kills consistently. Also, they do have areas or the option for them if they decide they want to affect large groups a bit more meaning more RP. Not to mention the fact that its easier to pick off targets from 30 range than it is a 0 - 2.5.

[ArchmagusArcana]

This may be so, but archers in general are doing a bulk of the fighting meaning more RP per kill and more kills consistently. Also, they do have areas or the option for them if they decide they want to affect large groups a bit more meaning more RP. Not to mention the fact that its easier to pick off targets from 30 range than it is a 0 - 2.5.

Exactly my point. Lots of warrior rp come from areas and auras, same for conju and to some extant locks. Most rp from archers are from direct confrontations, which to my mind means that they have to do more work to earn them, which makes their much higher than average rp even more telling than simply having a higher rp average.

[backe]

There are a lot of good points, but you guys are also forgetting how survival plays in this scenario. Given the large amount of survival skills that archers have (SotW, LP, Escapist, etc.) and given that you don't collect RP for targeted damage once you've died, its pretty easy to see why the RP board is as it is.

Another point, is that most hunters get the vast majority of their RP from hunting. Previously, barbs could help other classes hunt with friendly hunters, however this is not the case anymore. Even a max-speed skilled marks has a difficult time keeping up with a hunter without pulsing SotW between mobility occasionally. And of course, we have the current metric of only "hunters can catch hunters". As such, hunters are pretty much the only subclass that can efficiently hunt.

At fort wars, marksmen are easily the king of the hill when it comes to defense and survivability. Couple that with the ability to do damage at extreme range, then it becomes obvious that you have a class that can be well defended yet still snipe off damaged targets with relative safety.

All in all, the player base and their choices reflect the state of the game, and right now its pretty obvious what that is...

[Recoil]

I can tell you without any script that warriors are in terrible state now...:fury:

[Foolou]

It's rather the knights that are problematic, there is no significant difference between barbarians, conjurers and warlocks.

Apart from that I noticed that the general level of gained rp per day is very different from server to server, so I will take that into account. Once I have a bigger set of data, maybe after collecting the daily stats for two weeks, I will post some more detailed statistics.

I really hope that NGD does stuff like this, too. They have access to much more data, like playtime of the chars, their items, the skills they use, their damage on the other classes, their average time of survival in battle. It should be fairly easy to identify the problems that lead to the weak performance of knights.
There are so many correlations and cross correlations that could be taken into account, and maybe the changes that are needed to balance out things are just tiny adjustments.

[Dupa_z_Zasady]

Somehow hunters and marksman become a very deadly tool in the hand of the very ambitious players,

Please! Shooting in the back of low HP grinder is ambitious? You want me to die laughing?

[Eaten]

Exactly my point. Lots of warrior rp come from areas and auras, same for conju and to some extant locks. Most rp from archers are from direct confrontations, which to my mind means that they have to do more work to earn them, which makes their much higher than average rp even more telling than simply having a higher rp average.

I don't think its them having to do more work so much as they have the ability continuously work. Warriors spend much of their time moving to and fro. Archers have an easier time spanning that gap and can constantly fight, where as warriors have to get to the target and then start fighting, if things go badly they have to attempt to run back to conjurers and the cover of the archers. So, by this, ranged characters do a bulk of the work in war because its easier for them to do so.

[Foolou]

This is the result of a comparision of the data from sept 8th and sept 14th, boiled down to a relative rp weight of the classes in %:

Barbarian: 98
Knight: 66
Hunter: 122
Marksman: 114
Conjurer: 101
Warlock: 96

The intention is to again show that there is a fundamental difference between knights and barbarians in pvp.
The upcoming changes to the combat system will probably lead to a situation where barbarians are the only ones at the end of the food chain.
I hope there will be further modifications to balance out barbarians and knights.

[ArchmagusArcana]

Its not a balance issue, its an issue of knights dont do damage, which they shouldnt as they have all this armor and a broken block to tank damage. The problem lies in the fact that they are hard to kill and dont do damage, low priority target, therefore they dont get as much rp. Its got nothing to do with balance, its just poor mechanics.

Also, for your percentages to have any bearing, it should be a total of 100% only, those numbers have no meaning.

Edit: This is not a PvP game, its RvR, and if you look, warriors are at the very bottom of the 'food chain' as it is, the whole class needs an overhaul, which it is getting in the next update (at least as far as sorting out some poor mechanics).

[Foolou]

You're right that rp-measures are probably not good to tell something about performance of a class in an rvr game, because it almost never is a pure one vs one situation. Of course taking the rp average of one class and comparing it to another is neglecting most of the effects.

But it could be used as a correction factor for comparing your rp to someone elses, for example if you're a knight and you're comparing yourself to a hunter, you can almost double your rp to be 'comparable'.

Also, for your percentages to have any bearing, it should be a total of 100% only, those numbers have no meaning.

Well as I said I calculated a relative rp weight of the classes (not the total), and I'm using the % in the way it is defined : 1% = 1/100 (one per cent). So instead of writing a number like 0.96 it's 96%, nothing wrong about that. Of course you could divide all the values by 6, the relation would still be the same.
I did it this way to have the majority of the classes at a factor of almost 1 (barbs, conjurers and warlocks) and to see the deviation from that.