The Oracle of Death

[maradine]

Originally Posted by Chaff

Tanks again

I see what you did there.

Lightnings:



Original: http://bit.ly/16fKTHU

MBT reprinted as courtesy:



Glad I doublechecked all the Lightning data against the API - this is one of the few places where SOE breaks the NC,TR,VS ordering in the item ids. I'm actually a little curious if that's an error.

[mrmrmrj]

It has been written by others but I do think it is important to not make OP/UP judgements on every weapon based on KPU. This is because KPU will favor a weapon that is 3rd best at all ranges over any weapon that is best at one range and bad at others. In this sense, the KPU stat will tend to make specialized weapons look worse.

Looking at the Carbine run, the fact that the guns with the potential for grenade launcher attachments have higher KPUs than the others supports this interpretation.

Think for a minute about the extreme example. Let's say a game's guns are balanced around 3 distances to keep it simple: 10m, 50m and 100m. Would you rather use a gun that was pretty good at all three ranges or one that was very good at one range and terrible at the other two? In PS2, sometimes you know you want the 10m and sometime you know you want the 100m, but there are many times when having only one or the other means you die if the enemy has the correct one and you don't. This is why the SV-88 looks so good.

Lastly, I want to address the Striker. It looks completely balanced based on KPU, but KPU ignores the battlefield reality of this weapon. If 3 TR target a single vehicle with the Strikers, that vehicle is dead in the time it takes only one of them to fire because they can all fire simultaneously. If 3 Annihilators all target the same vehicle simultaneously, that vehicle (unless it is a Flash) is not going to die. The reload time will give the target the chance to escape.

The Striker gives the TR (and I do play TR sometimes so I have seen it from both sides) the power to quickly eliminate any single vehicle from the battlefield in one magazine with combined fire. This means that the Striker is best suited to killing the fastest, most maneuverable vehicles because of the lack of the need to reload. Combine that reality with the fact that the TR Harasser has the best AI option and we have the reality of PS2: the faction with the best anti-Harasser weapon has the best Harasser.

90% of my time is as NC. My outfit is good at focusing on one target with Annhilators. Anyone knows that 3 Annhilators will not kill most Harassers. Anyone who has played TR knows that 3 full Striker clips will kill any Harasser. The fact that each Striker missile does less damage individually than each Annhilator or Phoenix does not balance the weapon because 90% of the time I fire a Striker, I hit with all 5.

All 5 should do more damage than 1 Annhilator or 1 Phoenix. I do not disagree with that, but the spread needs to shrink.

[maradine]

Just had a quick convo with SOE API team - all reported hourly data is aggregate from launch.

[Shiladie]

Very interesting tank stats, keeping in mind we're just looking at 37 hours of data here.

Things to take from it:
- TR pull less skyguards, probably due to strikers filling the role.
- VS uses their MBT ES secondaries more than TR/NC, who make-up the gap with their halberd and basilisk use
- Prowlers are being pulled ~1.35x as much as other MBTs, with a slightly higher KPU at all skill levels
- Most weapons have a fairly straight forward KPU increase as BR goes up, but the Prowler HE is an outlier. the KPU spread of the prowler HE is 22.24 (7.4x), the highest spread between different BRs of any tank weapon, by a decent margin too.
- Saron is the most commonly used MBT secondary
- TR does not like their HE lightnings, while NC really likes theirs, doing extremely well with them, clocking in second behind Prowler HE for KPU.

Overall I wouldn't say the prowler is OP or anything, but it is definitely the high end outlier, at least with the numbers shown.
This is causing TR to neglect their non-viper lightnings though, with a poor showing there.
That and the enforcer C85 modified is a joke weapon.

[GS Ourous]

Excellent.

Thankyou for the data, I will use this to continue my argument with another Lightning driver regarding who is superior.

[Plaqueis]

I knew the Prowler was easily the best mbt, but i didn't except the margin to be that huge... same goes for Striker. Popularity effect?

[Chewy]

Originally Posted by Plaqueis

I knew the Prowler was easily the best mbt, but i didn't except the margin to be that huge... same goes for Striker. Popularity effect?

I think that the shear over use of the Striker is the reason for such good Prowler KPUs.

Go back a few pages and look at my opinions on the lock-ons and I said that the Striker is so used and popular that it has to fight itself for kills. If an AV weapon is fighting itself then that frees up or even forces its teammates vehicles to farm infantry as enemy vehicles have to deal with the Strikers first or they can't last long enough to engage other vehicles or get infantry farm.

Without long lasting enemy vehicles then your vehicles should be able to get a lot of farm as they don't have to choose to deal with them and give up getting easy infantry kills or risk getting shelled in the ass. Even if Strikers don't kill enemy vehicles, those vehicles MUST run to recharge smoke/shield/flares or die from the unstoppable rain of missiles. Either way a single TR HA can remove a vehicle by forcing a counter measure or just dealing a good chunk of damage.

[Falcon_br]

Great oracle of death:
How about all pistols data?
I know all NS pistols will be the most in use, because they are common pool, but I would like to know about the other secondary pistols, I use the TX-2, but I think I might be the only one doing it!

[Plaqueis]

Originally Posted by Chewy

I think that the shear over use of the Striker is the reason for such good Prowler KPUs.

Go back a few pages and look at my opinions on the lock-ons and I said that the Striker is so used and popular that it has to fight itself for kills. If an AV weapon is fighting itself then that frees up or even forces its teammates vehicles to farm infantry as enemy vehicles have to deal with the Strikers first or they can't last long enough to engage other vehicles or get infantry farm.

Without long lasting enemy vehicles then your vehicles should be able to get a lot of farm as they don't have to choose to deal with them and give up getting easy infantry kills or risk getting shelled in the ass. Even if Strikers don't kill enemy vehicles, those vehicles MUST run to recharge smoke/shield/flares or die from the unstoppable rain of missiles. Either way a single TR HA can remove a vehicle by forcing a counter measure or just dealing a good chunk of damage.

Yeah well, no matter what's the actual cause, some of these gaps are freaking huge.. and yet not really surprising.

Harasser Weapons

G20 Vulcan-H
id 6121
4,777 kills
815 unique killers
5.86kpu

Saron HRB-H
id 6126
1,702 kills
402 unique killers
4.23kpu

Enforcer ML65-H
id 6123
958 kills
270 unique killers
3.55kpu

^My favorite, specially since the Saron got nerfed due excessive whining...

id 7528 NC05 Jackhammer
17,151 Kills
1,216 Unique Killers
14.10 kpu

id 7533 T7 Mini-Chaingun
37,004 Kills
2,600 Unique Killers
14.23 kpu

id 7540 Lasher X2
16,144 Kills
1,656 Unique Killers
9.75 kpu

^Wasn't there a huge whine-fest about the spooltime of T7..? Doesn't seem to have too much effect on it...

id 3700 Titan 150 HEAT

HEAT (or equiv.)
===

id 4000 P2-120 HEAT
50,606 Kills
6,278 Unique Killers
8.06 kpu

id 3400 Supernova PC
28,452 Kills
4,677 Unique Killers
6.08 kpu

id 3700 Titan 150 HEAT
26,377 Kills
4,997 Unique Killers
5.28 kpu



AP (or equiv.)
====

id 4008 P2-120 AP
11,429 Kills
1,123 Unique Killers
10.18 kpu

id 3460 Supernova FPC
6,959 Kills
823 Unique Killers
8.46 kpu

id 3730 Titan-150 AP
9,839 Kills
1,242 Unique Killers
7.92 kpu



HE (or equiv.)
===

id 4009 P2-120 HE
10,239 Kills
606 Unique Killers
16.89 kpu

id 3461 Supernova VPC
7,386 Kills
612 Unique Killers
12.07 kpu

id 3731 Titan-150 HE
4,054 Kills
435 Unique Killers
9.32 kpu

^Case already discussed...

Seems that in most cases, NC and VS stuff is more or less balanced and TR is on class of its own for whatever reason. Only things different so far are what, lmg's and maxes, and be sure theres a TR player whining about them being too OP... lolz.

But it is what it is, i don't care. I trust the developers will take action if they see fit. Meanwhile, i'll be just enjoying the game.

[maradine]

Originally Posted by Plaqueis

But it is what it is, i don't care. I trust the developers will take action if they see fit. Meanwhile, i'll be just enjoying the game.

Amen.

True story - I literally fell asleep breaking out the rest of the Harasser data update - that should be up this morning. I'll try to get in the full pistol spread, too.

I'm having some interesting discussions about how real-time play time might be tracked, since so many feel this is an important metric (spoiler alert: still think its going to say the exact same thing, ala the MAX comparison in the early pages where we compared kpu metrics to kph metrics).

This is largely complicated by the fact that SOE only reports playtime aggregate. When you see kph on ps2-stats, you're seeing the aggregate playtime reported since PS2 launched their tracking. A possible solution is to get the save state data as it crosses the wire, just like kills, throw the results into memcached, wait for a matching later update, and throw the diff back into the Oracle. Messier still is that fact that long player sessions can push the save state update much farther out in time than the kills themselves, splitting some playtime recording across the time line from the kills associated with it.

Finally, all of this needs to fit inside the hardware constraints I set up for the project. It's so easy when you're developing to have hardware solve your shitty optimization problems. I don't need any more four figure monthly AWS bills this year (long story).

So - more to come! I'm getting less confident about being able to avoid a nuke'and'pave in the next month, but I'll try to hold it down so we get a full GU that I can aggregate and copy off before I do so.

Those of you who like utterly boring debug information can point your IRC clients into irc.slashnet.org, #planetside9, where the Oracle quietly murmurs into the darkness about once every 20 seconds. You can also get yourself a slot in #planetside2, which will be where the production version will live. Those FKPK guys were just leaving- I promise.

[Plaqueis]

Originally Posted by maradine

True story - I literally fell asleep breaking out the rest of the Harasser data update - that should be up this morning.

Can't wait.. i just lost twice to a Vulcan-Harrasser in my Skyguard, no chance whatsoever for me.. i barely scratched it with my approx 50 40mm hits on it, his gunner sprayed me for maybe 2 seconds and i blew up. I haven't tried these suicideruns with my Saron-Harrasser, so i have no clue if it's similar.

EDIT: Gotta say though, if i'm soloing without my outfit, i choose not to even try to fight TR if there's an option to go against NC. 'Funfactor' just isn't there.

[EVILPIG]

Dear OoD, this thread is too long to read. I read 6 pages, skimmed the rest. If you have not already done so, for shiits and giggles, please post the overall KPU of the 3 empires.

[maradine]

Originally Posted by EVILPIG

Dear OoD, this thread is too long to read. I read 6 pages, skimmed the rest. If you have not already done so, for shiits and giggles, please post the overall KPU of the 3 empires.

I've been avoiding that one out of loving kindness for my DB engine, but fuck it - running.

[ChipMHazard]

Originally Posted by Plaqueis

I knew the Prowler was easily the best mbt, but i didn't except the margin to be that huge... same goes for Striker. Popularity effect?

The Prowler is the more popular choice against infantry and that's where the difference in kills comes from. The Prowler isn't necessarily the best tank at being a tank when going up against other tanks or enemy armour in general.
The Striker is the best lock-on weapon in the game when it comes to damage output, the TR doesn't need to use any other heavy AT weapon. VS/NS have to turn to other weapons if they want a lock-on weapon. Basicly you're seeing a split of kills over more weapons, while the TR don't have to split their kills. Also note that the Striker is effective against both air and ground.

So yes, popularity.

[maradine]

Originally Posted by ChipMHazard

The Prowler is the more popular choice against infantry and that's where the difference in kills comes from. The Prowler isn't necessarily the best tank at being a tank when going up against other tanks or enemy armour in general. So yes, popularity.

So the implication is that, even for Prowler pilots who pull AP, they're still farming infantry?