Evaluating differences in sustain between factions - LOTD 12/05/24

[leftayparxoun]

Hello everyone,

This is not a exactly a Balance Proposal post, more so a ''I-found-something-interesting'' post.
If this leads to the balance team gaining new insights from what you will read here, then so be it.
The methodology that I ended up using can also be applied to other (new and old) skirmishes to gather similar data, although I would caution people who end up using it to make sure to compare at least decently equal forces. I tried to prove that this is the case in this present analysis.

Without further ado, here is what I've worked on: https://docs.google.com/document/d/1pyq ... sp=sharing

It first goes through the theory, then presents the results from the data of this LOTD and finally contains my personal interpretation of those results.
As I also write in the conclusion, if you can come up with any additional/alternative interpretations of the results, by all means, share it with us here.

For those who don't want to spend time reading through the entire document, here is the briefest TLDR I could make:

- You can quantify Recovery (=sustain through heals) via the killboard Kill Damage stats.
- This works for both factions and you can calculate a time-averaged value for each of them for the entire skirmish.
- In this LOTD both factions were very similar in organization and in gear/renown ranks, so it makes for a good dataset.
- From the results: Destro can seemingly sustain much better than Order (despite Order being designed to do so better)

According to my Interpretations, this can be explained:

- By Order blobbing way more than Destro AND by being severely unbalanced (unlikely)
- Or by only losing the fights where the odds were stacked against them. For the rest of the fights, various small balance factors (that you can read in the document) were adding up, favouring Order in LOTD warfare and which could potentially be addressed (more likely)

Hope we can have a constructive discussion about these findings

Edit: Adding an additional analysis of a balanced City fight from the next day: https://docs.google.com/document/d/1tqj ... sp=sharing

[Hazmy]

Moved to "Suggestions & Feedback" Forums, it was in the wrong section.

[Akalukz]

Just a brief comment, i think that is exactly how the game was designed. Order stronger in small/even fights, destro stronger in the zerg. Heroic Last stands etc.

Great anyalysis, i actually read it But i didn't see a class breakdown? Which could greatly skew your results. If order was primarily made up of Eng / SW damage will be high while kills would still be low. Same is true of Destro with a lot of mara / Magus.

Can you get class data?

Edit:

Additinally i don't think you lend enough credit to the LoTD Buff/Debuff it is local to the area so can greatly skew certain engagments etc, maybe your analysis of the skirmishes could be broken down into smaller increments?

[Absinth]

Just a brief comment, i think that is exactly how the game was designed. Order stronger in small/even fights, destro stronger in the zerg. Heroic Last stands etc.

Great anyalysis, i actually read it But i didn't see a class breakdown? Which could greatly skew your results. If order was primarily made up of Eng / SW damage will be high while kills would still be low. Same is true of Destro with a lot of mara / Magus.

Can you get class data?

Edit:

Additinally i don't think you lend enough credit to the LoTD Buff/Debuff it is local to the area so can greatly skew certain engagments etc, maybe your analysis of the skirmishes could be broken down into smaller increments?

The skirmish system we currently have is very bad and makes small scale analysis way harder to do as it likes to group small fights into big fights.

[leftayparxoun]

Just a brief comment, i think that is exactly how the game was designed. Order stronger in small/even fights, destro stronger in the zerg. Heroic Last stands etc.

Great anyalysis, i actually read it But i didn't see a class breakdown? Which could greatly skew your results. If order was primarily made up of Eng / SW damage will be high while kills would still be low. Same is true of Destro with a lot of mara / Magus.

Can you get class data?

Edit:

Additinally i don't think you lend enough credit to the LoTD Buff/Debuff it is local to the area so can greatly skew certain engagments etc, maybe your analysis of the skirmishes could be broken down into smaller increments?

Thanks for the reply. I added the class breakdown as requested. I wouldn't say it changes anything about the results.

TLDR:
- Tank distribution is similar between factions, just IB is not preferred.
- DPS distribution is very slightly in favor of Order for highly geared dps (especially Bright Wizards vs Sorcs). Destro actually had less number and % of ''bad classes'' (WE, Magus) compared to Order (WH, Engi). But even for those classes Destro had better gear on them.
- Healers are hard to make a judgement (due to the existence of dps Healers in 6mans) but it seems like Order had fewer but overall better geared with the exception of RP vs Zealot.

I actually address the LOTD debuff in the subsection ''Can LOTD debuff explain the results?'' in the Interpretation section. There I show how the numbers can be explained by mainly by debuff but that would:
1) Require drastic differences in blobbing between factions which we did not observe
2) Indicate that there are some huge innate power gaps between individual Destro and Order warbands (which imo isn't really observed in oRvR)

Unfortunately this method only works when adding everyone's kill damage in each faction and as such cannot be used for individual conflicts in the skirmish.
It could, however, find application in smaller RvR/scenario skirmishes at the risk of not meeting the equal-organization and equal-gear conditions.

[nocturnalguest]

Awesome read. Thanks! I will allow myself a brief comment tho.
Thats very interesting approach and even with superficial showcase (being obviously limited by alot of factors and constraints of skirmish data).
A big fat kudos that you are not taking responsibility to make "conclusions" out of this. Very noble, fair and rightful.
Ill think a bit and study this in details to try to give some feedback.

[Phantasm]

Where in those numbers and calculations is the player factor? Is it possible to compare the kill/death ratios of characters with at least a 78+RR and compare them there?

[leftayparxoun]

Where in those numbers and calculations is the player factor? Is it possible to compare the kill/death ratios of characters with at least a 78+RR and compare them there?

To put it simply: That factor isn't really relevant

KDR of players is an available metric but it cannot fit in the proposed methodology due to killboard also being ignorant regarding Kill Damage vs specific players. Instead it sums up all the kill damage performed (by each player). This methodology takes advantage of the following relation to circumvent that:

"The total Kill Damage done by a faction is the total number of people who died on the enemy faction times their mean health plus the healing they recieved during the 100-to-0 process."

Due to that collective nature, it cannot tell us more about individual performance. One "easy" improvement would be to calculate the HP total of each player who died (from killboard/armory) and multiply that with their deaths to get a better estimate of the HP total/average HP of the faction. This option, despite skipping the HP assumption, entails the risk of:
a) Them having changed their equipment/build since then
b) ignoring if they are getting more Wounds from linis/RR points/teammate buffs

In this analysis, the scope of LOTD makes looking into individual deaths and interpreting what happened extemely difficult. If the same methodology was applied in a smaller skirmish (e.g. City fight, or scenario) with a less sizeable amount of deaths on each side, then you could make a case for looking into individual kills.
For LOTD, averaging values over the entire faction and over the time of the entire skirmish makes more sense. Especially when both factions are quite similar in degree of organization and gear/renown.

[Gakin]

I think your assumption of "Similar Degree of Organization" is wrong, you cant just count number of guilds, order had more good wbs in my opinion. Good point on SM healing, its just broken in LOTD.

[Ashoris]

Hi,

I applaud you for the detailed analysis.
The only thing that i want to mention is that your assumptions of how the "REALM" is designed and works is something you should review to make sure it does not screw your conclusion.

You found some unique order skills that are not cloned to destru side but ignored some key aspects in class design.
I would highly suggest to also look at how classes are designed in general and how much sustain is inbuilt. Perfect examples would be WE/MSH/Mara vs WH/aSW/WL
or AM vs Shaman ....

To write that order is built around sustain is at least questionable.

TO be Fair RoR made it more difficult to have a single identity because they removed a LOT of the core mechanics of the realms and built in several changes without a design philosophy ... (order was crit stacking king while destru was a morale pump nightmare )