Lots of Numbers: ORVR last 10 days

[juzziex]

You can't trust the top scores on the KB. They have either WB's supporting them or are SC goblins who have multiple tanks and healers in a pre-made supporting 1 DPS pumping their killboard score.

Well you can't trust stats in this game in general. Good damage comes from good support and/or spamming fluff AOE. Premades can easily farm pug. Certain people love to just run away from hard fights and leave their group to die to pad their stats. Solo kills are scuffed by heals/guards not getting counted in the solo tracking etc and a lot of solo players actively run away from people they find hard to beat instead of giving it a go anyway and learning how to win.

Still its interesting to see as most people would still agree that the stat leaders here are generally thought of as the best players, regardless of it not being completely accurate.

[Acidic]

Was thinking how to get the numbers standardized.

One thought was to try band the results based on rr 83+ (almost certain bis), 78-82(sov), ….
This type of banding can normalize the groups , would be interesting also just to see the numbers to see how rr on order vs destro lines up.
Sorry too lazy to do the actual work.

So could be interesting
RR group, class, count, total kills , total deaths, median kill, median deaths, average kills, average deaths, median k/d, average k/d


Use of median will reduce the outliers .

The use of the bands helps mitigate arguments that populations are different geared .., gives insight to how different factions are built up, and also see if there is performance change on gearing. (I know Mara pre 78 and after 78 was noticeable change)

Might even be worth grouping into time zones, to see the difference , based on server time, 03-09, 09-15, 15-20, 20-23, 23-03 (just example of how I see different pop changes)

Normally I would have a meaning of what the stat should be showing before producing the numbers so as to avoid bias interpretation.

[Vaul]

rvrstats website mostly does this. e.g. RvR week so far (week 36) RR80+, median KD ratio, timezone 6-11pm UTC (~ EU prime)

[Acidic]

rvrstats website mostly does this. e.g. RvR week so far (week 36) RR80+, median KD ratio, timezone 6-11pm UTC (~ EU prime)

I know, it’s a good site. All I was trying to do was encourage the op to put stats out that are inteptable rather than just dumping numbers.

[githappens]

Such stats are worthless, because the only thing they indicate is activity of leaders like Phature compared to ones like Ywhi. You can see huge gap between those, despite both being BiS and having guard and heals 100% of time.
You can find same for every class not only slayers, e.g. average SW stats vs Zeromancer or Fenryl.

[Evilspinnre]

Such stats are worthless, because the only thing they indicate is activity of leaders like Phature compared to ones like Ywhi. You can see huge gap between those, despite both being BiS and having guard and heals 100% of time.
You can find same for every class not only slayers, e.g. average SW stats vs Zeromancer or Fenryl.

One of them actively leads RvR wbs on a regular basis while the other dies occasionally in city and spams scs?

[githappens]

Such stats are worthless, because the only thing they indicate is activity of leaders like Phature compared to ones like Ywhi. You can see huge gap between those, despite both being BiS and having guard and heals 100% of time.
You can find same for every class not only slayers, e.g. average SW stats vs Zeromancer or Fenryl.

One of them actively leads RvR wbs on a regular basis while the other dies occasionally in city and spams scs?

I find it poor argument. As Ywhi used to lead on slayer in oRvR with similar results in old days.
My point was, comparing average class stats is pointless, because those numbers are affected by indivitual skill and warbands (renown, gear, setup etc all matter) those players play in.
If you want, you can compare to Nasilni, if you want compare to someone who plays alot of oRvR on order in warbands.

[Acidic]

Such stats are worthless, because the only thing they indicate is activity of leaders like Phature compared to ones like Ywhi. You can see huge gap between those, despite both being BiS and having guard and heals 100% of time.
You can find same for every class not only slayers, e.g. average SW stats vs Zeromancer or Fenryl.

One of them actively leads RvR wbs on a regular basis while the other dies occasionally in city and spams scs?

I find it poor argument. As Ywhi used to lead on slayer in oRvR with similar results in old days.
My point was, comparing average class stats is pointless, because those numbers are affected by indivitual skill and warbands (renown, gear, setup etc all matter) those players play in.
If you want, you can compare to Nasilni, if you want compare to someone who plays alot of oRvR on order in warbands.

I do agree that comparing 1:1 classes is a problem, the stats do not nessasaraly say anything about the class but says things about the total situation, group, player, population….

The only way to be able to genuinely compare classes is to somehow to standardize the sample.
It’s clear enough to see this when looking at choppa/slayer over the different time spans avalable on rvr stats.
The problem is to an extent the population size. If you look at eu prime time there is a clear impact on choppa pop from organized warbands, this impacts the whole day choppa stats due amount of activity at those times.
It was semi possible to look at stats in season 2 (think it was this one) as it was possible to remove certain factors and there was much less “fixing and trading” matches.
So in short it is extremely difficult to do stats in a meaningful manner when you are unable to have comparable groups.

[gersy]

lol witch hunter very low K/D ratio but still getting giga nerfed

[Phantasm]

I do agree that comparing 1:1 classes is a problem, the stats do not nessasaraly say anything about the class but says things about the total situation, group, player, population….

The only way to be able to genuinely compare classes is to somehow to standardize the sample.
It’s clear enough to see this when looking at choppa/slayer over the different time spans avalable on rvr stats.
The problem is to an extent the population size. If you look at eu prime time there is a clear impact on choppa pop from organized warbands, this impacts the whole day choppa stats due amount of activity at those times.
It was semi possible to look at stats in season 2 (think it was this one) as it was possible to remove certain factors and there was much less “fixing and trading” matches.
So in short it is extremely difficult to do stats in a meaningful manner when you are unable to have comparable groups.

In last Dev Update they shared something called Automated Test System, which means they can test every combo of skills and compare results. I guess its great tool to compare dmg potential of every class and let do tweaks here and there to balance them out.

There is an analogy to cars made by manufacturers.

• Just as car manufacturers aim to pass crash tests and emissions standards, developers aim to meet class balance benchmarks

• Just as car owners can customize their vehicles with aftermarket parts, RoR players can tailor their class setups with gear, tactics, and skill rotations.

These choices can significantly affect how a class performs in the heat of battle. Mike might opt for survivability and control, while Sarah could focus on maximizing DPS.