The Oracle of Death

[Chaff]

.
....oh Great and Powerful Wizard of OZacle....

Have you thought about looking at some of the data from the other side of things ?

Such as, sort deaths per MBT spawned based on BR,...... ?
Does SOE provide data on time until death/loss of a vehicle after it is spawned at a vehicle pad oputside the WG (or leaves the WG after being spawned there) ?

There are many many reasons and variables that determine how long a vehicle survives after being spawned. Still, if the data is there it may help shed a little more light on the discussion.

The longer an individual player, or weapon platform, survives in the battlefield the more opportunities to build higher KPU numbers.

Did I see earlier that you had done overall KPU per Empire, and they were all around 41 KPU ? Did I actually read that earlier ? If overall KPU appears to be in balance, does that mean every other metric is essentially us confusing the issue with another unique serving of minutiae ?

KPU seems like it measures "upside" of the BEST operators of said gun, weapon, vehicle,......

Perhpas. Perhaps, if we measure Death of "X" vehicle configuration/platform at different BR tiers, we'll learn more about how a particular Learning Curve (skill level) may be skewing numbers. TIME UNTIL DEATH seems vital to discerning effectiveness of a weapon, armor, loadout, playstyle,......

This sounds incredibly daunting. Even if the numbers are there, time until death needs to be measured vs kills. How many times are you being shot at per minute (on average). How much are you firing on a target per minute ? When "X" player is in a "hot zone", we need to know these numbers on at least a thumbnail level.

Some of the non-engagement time is a vital part of the puzzle. A sniper, tank gunner, 2ndary gunner, ESF, Dalton gunner, .... all need time to position themselves to avoid damage yet also acquire targets from the most advantageous position(s) possible. I'm not sure we can come up with the proper data filtering tool to get helpul meaurements of what makes one weapon (or even one tactic vs another) better than the wide range of other options available to players.

Noob-fiiendly, Cert/Resource friendly weapons or platforms may have lower KPU simply because the talent pool using X weapon/platform.

High-SKILL (and/or High Cert Level) based weapons may disproportionately have skewed KPU because unskilled players aren't prevelent enought to lower overall KPU percentages (gunner efficiency, .....errrrr....SKILL)).
.

[maradine]

Originally Posted by Chaff

.
Noob-fiiendly, Cert/Resource friendly weapons or platforms may have lower KPU simply becasue the talent pool using X weapon/platform.

High-SKILL based weapons may disproportionately have skewed KPU because unskilled players aren't prevelent enought to lower overall KPU percentages (gunner efficiency, .....errrrr....SKILL)).
.

This is the primary reason I do the quartiles - you can see how much the wielder's experience (a not perfect, but decent, proxy for skill) effects his production. You can even go so far as to infer that weapons with very high variances between Q1 and Q4 (most infantry weapons, many vehicle weapons) are High Skillcap Weapons (or at least, hyper-potent upgrade-carrying weapons) and that weapons with low or no variance (lock-on launchers) are Low Skillcap Weapons. I've been contemplating breaking that out. Interesting things also happen when you divide a weapon's KPU by its AVGBR and a constant. More on that in later sheets.

To your other point, I don't think deaths will be happening soon, but vehicle destruction is next up, and there are some interesting inferences to discuss.

[EVILPIG]

I wouldn't get too specific. While these numbers tell a part of the story, they don't tell everything and BR says something, but not much, as many high BR players have lower BR alts and they perform just as well with them.

[maradine]

They do, and you can see that effect. To say that it is more common than not, however, would be a massive assumption that I don't feel is warranted.

In any event, we'll never know for sure, as that data will explicitly not be published*.



*to us

[EVILPIG]

Did I miss the overall KPU per empire?

[maradine]

Originally Posted by EVILPIG

Did I miss the overall KPU per empire?

Yeah, it's back there - lemme pull it up.

Post #186

[EVILPIG]

Originally Posted by maradine

My guess is that the average TR Lightning HE gunner is 10 BR less than the average NC Lightning HE gunner.

It could also be that the id numbers got munged - feel free to double check that I've applied the right faction to the right weapon id.

Results coming in on EvilPig's query, will be edited in here:

Global Death Stream, project inception to end of GU13

Vanu Soverignty
1,537,990 Kills
37,407 Unique Killers
41.1150 kpu
40.7833 avgBR

New Conglomerate
1,418,202 Kills
38, 328 Unique Killers
37.0017 kpu
38.8333 avgBR

Terran Republic
1,556,638 Kills
38,566 Unique Killers
40.3630 kpu
39.6361 avgBR

That's a <2% margin between #1 and #2, and a 11% margin between #1 and #3.

Thanks!! Interesting how little, yet how much it says! lol

Most interesting is how close each empire is in population.

[DviddLeff]

Any chance we could get stats that show the overall results for all of a particular type of vehicle, eg average KPU for ESFs, MBTs, etc?

Apologies if it is already posted.

[maradine]

Bad news - came across an API bug in the vehicle reporting that won't be fixed until the next GU. So I'm not going to be able to marry vehicle destruction to death records and find out, for instance, how many tankers are going down with their ship. On the other hand, something else to look forward to in GU15.

[maradine]

Scratch that - next hotfix!

[Sledgecrushr]

O great oracle, what are the top 5 weapons overall?

[maradine]

An interesting question. Would you like late GU13 or early GU14 comparisons?

[Yetskii]

So Oracle, I have a small question....

What is the number one killer of Sunderers?

[maradine]

Originally Posted by Yetskii

So Oracle, I have a small question....

What is the number one killer of Sunderers?

Ooo, I like that one. Not sure if I can get it without the next API bugfix, but I'll see what I can do.

[Sledgecrushr]

Originally Posted by maradine

An interesting question. Would you like late GU13 or early GU14 comparisons?

Early gu 14 would be the most relevent, most gracious Oracle.