DUST Search - Sticky:[Data Insights] Dropsuit KDR

CCP Rattati wrote:

Adipem Nothi wrote:

Tesfa Alem wrote:

1. No matter how many were sold if its was OP as people claim it should have the highest KD. It isn't.

2. Also, we should absolutley ignore PC data when balancing for everyone ... PC is the worst metric to go by. Balance for that and break the game for everyone.

Point 1. That isn't necessarily true. If this report were restricted it to the top X% of performers, then you would be more correct. We're looking at averages here, not distributions. To illustrate (or at least attempt to Lol

Adipem Nothi wrote:

Assumption
Right now, 2000 vets are running Prototype Assaults. 500 are running Assault gk.0. 1,500 are running Assault mk.0. Of the 500 running Assault gk.0, 250 are fair shooters and 250 are good shooters. Of the 500 running Assault mk.0, 500 are bad, 500 are fair and 500 are good.

Count - Skill - KDR
500 - Bad - 2.0
500 - Fair - 3.0
500 - Good - 5.0
Average mk.0 KDR ~ 3.3

Count - Skill - KDR
0 - Bad - N/A
250 - Fair - 2.5
250 - Good - 4.1
Average gk.0 KDR ~ 3.3

Question
If this scenario were true, could we cite it as indication that Assault gk.0 and Assault mk.0 are perfectly in balance?

Point 2. I'm suggesting we check our expectations against all available datasources before ruling whether or not something is balanced; that's quite different than balancing for PC. If our expectation is that MN Assaults are fine but PC usage rates, PC efficiency rates, forum complaints and pub usage rates all point to the contrary, would it not be prudent to at least challenge our expectation?

To be honest, this is of course the right thing to do, just like with weapons, KDR plus use and actual effectiveness. I am trying to find the right combination of a key performance indicator (kpi) that just tells us whether or not something is relatively overpowered.

If I told you all these are the same, except 50% kills are performed by Minsaults, we would have a problem. Just like with the Rail Rifle before. But if the Kills are split within the Assault Role fairly equally, and the KDR is fairly equal, then I would be less concerned.

That's why I am running an experiment whether the right formula is K/S*relativeKill% for weapons. That quickly identifies weapons that kill a lot and have a high KDR, and isolates cases where KDR is good, but Kills are low, and thirdly, low KDR, low Kill weapons stand out like a sore thumb.


Author	Thread Statistics \| Show CCP posts - 5 post(s)
Tesfa Alem Death by Disassociation 1071	Posted - 2015.05.12 16:00:00 - [1] - Quote Vindication for my Min Assault. Suck it Nerds. On par with the gallente, and the caldari assaul has a better KDR at ADV. Just edging out the caldari assault at proto. Not enough for data to back up all the 'MIN ASSAULT IS SOO OP' threads. My Minmatar Logi is its rightfull place, and I wear that terrible KD with Pride. After all, if your a logi running around shooting people your doing your job wrong. Also I'm glad that you rattati came up with this chart instead of rolling out the PC stats. Can we get a similar comparision for dropships and tanks? Redline for Thee, but no Redline for Me.
Tesfa Alem Death by Disassociation 1071	Posted - 2015.05.12 16:42:00 - [2] - Quote Markets are all about percieved value, just like in IRL trading. Everybody talks about the Min Assault being OP, so everybody buys it. The more people use it, the more it appears on the kill feed. The more it apears there, the more it plays into the popular idea that it's OP. So the ADV min assault is the most pupular suit purchased. It still has the 3rd ranked KD of all the ADV assault suits. No matter how many were sold if its was OP as people claim it should have the highest KD. It isn't. Also, we should absolutley ignore PC data when balancing for everyone. e.g. In PC sentinels are the suits with a massive KD disproportiante to the game. In the entire dust 514 community, Assaults have the highest KD. The freaking proto Caldari scout has a higher KD than Minmatar, gallente and Amarr prototype Sentinels. PC is the worst metric to go by. Balance for that and break the game for everyone. Redline for Thee, but no Redline for Me.
Tesfa Alem Death by Disassociation 1073	Posted - 2015.05.13 09:11:00 - [3] - Quote CCP Rattati wrote: Adipem Nothi wrote: Tesfa Alem wrote: 1. No matter how many were sold if its was OP as people claim it should have the highest KD. It isn't. 2. Also, we should absolutley ignore PC data when balancing for everyone ... PC is the worst metric to go by. Balance for that and break the game for everyone. Point 1. That isn't necessarily true. If this report were restricted it to the top X% of performers, then you would be more correct. We're looking at averages here, not distributions. To illustrate (or at least attempt to ): Adipem Nothi wrote: Assumption Right now, 2000 vets are running Prototype Assaults. 500 are running Assault gk.0. 1,500 are running Assault mk.0. Of the 500 running Assault gk.0, 250 are fair shooters and 250 are good shooters. Of the 500 running Assault mk.0, 500 are bad, 500 are fair and 500 are good. Count - Skill - KDR 500 - Bad - 2.0 500 - Fair - 3.0 500 - Good - 5.0 Average mk.0 KDR ~ 3.3 Count - Skill - KDR 0 - Bad - N/A 250 - Fair - 2.5 250 - Good - 4.1 Average gk.0 KDR ~ 3.3 Question If this scenario were true, could we cite it as indication that Assault gk.0 and Assault mk.0 are perfectly in balance? Point 2. I'm suggesting we check our expectations against all available datasources before ruling whether or not something is balanced; that's quite different than balancing for PC. If our expectation is that MN Assaults are fine but PC usage rates, PC efficiency rates, forum complaints and pub usage rates all point to the contrary, would it not be prudent to at least challenge our expectation? To be honest, this is of course the right thing to do, just like with weapons, KDR plus use and actual effectiveness. I am trying to find the right combination of a key performance indicator (kpi) that just tells us whether or not something is relatively overpowered. If I told you all these are the same, except 50% kills are performed by Minsaults, we would have a problem. Just like with the Rail Rifle before. But if the Kills are split within the Assault Role fairly equally, and the KDR is fairly equal, then I would be less concerned. That's why I am running an experiment whether the right formula is K/S*relativeKill% for weapons. That quickly identifies weapons that kill a lot and have a high KDR, and isolates cases where KDR is good, but Kills are low, and thirdly, low KDR, low Kill weapons stand out like a sore thumb. I agree that KDR isn't the sole metric to go by. At the same time i'm seeing a lot of people on these forums rejecting this data because of thier own personal opions, more or less along the lines of 'it killed me it must be OP', or 'its popular it must be OP'. All in all its looking at the data backward. Instead of 'here's the data what is OP and what isnt' we have 'the Minmatar Assault is OP! Where is the data to justify this? But- i've been complaining how could i have been wrong and the data be right? Must find another source' hence my 'SUCK IT NERDS' comment. Because there are few trhings i hate more than the second line of reasoning. Screw the emotional state of the forums and screw PC, that is no way to balance dust at all. Redline for Thee, but no Redline for Me.
Tesfa Alem Death by Disassociation 1075	Posted - 2015.05.13 21:43:00 - [4] - Quote Adipem Nothi wrote: What Tesfa really wants is for others to learn to be logical and reasonable, just like him. Yay more me. Redline for Thee, but no Redline for Me.