Stats, Stats, Stats! A year of AH2.

[Krusty]

Originally posted by Lusche
Most of them got the PERK states just because they are better killers. If they were like any other plane, their would be not much need for any perking at all

Without perk limits, their K/D would drop considerable for sure, but planes like Tempest would still be ahead of the crowd in K/D terms just because of their performance.

Sometimes that's true. Sometimes not. Most claim the Spit14 wouldn't see too much use (as it doesn't fly like a spit) if unperked. The Ta152 surely doesn't see much use. On the other hand, the Chog did, and the 4hog would, as would the tempest.

There is too wide a range. Some planes are perked to limit them, not because they are powerful, and some are perked because they are powerful. I've heard one comment from somebody at HTC (long ago) state the Spit14 is perked because it's a monster, with some reference to its extreme rate of climb. On the other hand we know the chog wasn't perked until it was horribly over-used (this back when hogs were NOT uber! Imagine now!)


So I think the high ratio is more about lack of deaths, rather than surplus of kills. Look at the 262. Ranked #2, but definitely NOT a super-lethal killer, like the tempest. Probably in reality it's in the lower 30s of the entire planeset.

[Lusche]

Originally posted by Krusty
. Look at the 262. Ranked #2, but definitely NOT a super-lethal killer, like the tempest. Probably in reality it's in the lower 30s of the entire planeset.

Lower 30s in which regard? Sorted by what?

[whiteman]

I was thinking the same about the F4U's. I know when i get in a perk ride I'm more careful take the time to climb out to a good alt to be above any fight. Even when i do loose a perk ride it's after multiple sorties in it of 3-4 kills a piece then i park it and go back to the -1A.

[hubsonfire]

The only things the 262 has going for it are its range, endurance, speed, acceleration, and armament. It is absolutely one of the least lethal planes in AH.

[Krusty]

Don't ever quote me on this, but ...


"What hub said!"

Lusche, just in an overall "effective killer" kind of ranking.

Try re-running the numbers with JUST kills, not kills+deaths. You'll find that they shift a lot. There are a lot more kills in the top3 than the others. Deaths can be from augering, second account killing, wing ripping for no fershlugginer reason, kill shootering, and many many non-combat-related reasons. Or just screwing around.


Just use kills and the outcome is a better representation on the chart.

[Adonai]

Originally posted by hubsonfire
The only things the 262 has going for it are its range, endurance, speed, acceleration, and armament. It is absolutely one of the least lethal planes in AH.

Most that fly the 262 know it takes skill to fly, where as tempest is less then HALF the perks of 262 and much easier to fly.  Alot of pilots ive talked to simply dont have the perks or do not want to spend 200 plus perks on it, since they can get same kills in a tempest and its cheaper. In regards Both planes are unique in their own way, and needs to be flown by skilled pilot or simply a "spitfire" could catch it and shoot it down.

[Lusche]

Originally posted by Krusty
Don't ever quote me on this, but ...


"What hub said!"

Lusche, just in an overall "effective killer" kind of ranking.

Try re-running the numbers with JUST kills, not kills+deaths. You'll find that they shift a lot. There are a lot more kills in the top3 than the others. Deaths can be from augering, second account killing, wing ripping for no fershlugginer reason, kill shootering, and many many non-combat-related reasons. Or just screwing around.


Just use kills and the outcome is a better representation on the chart.

No.

The "usage" chart (where kills and deaths are combined) was intendend to show just that, how much a plane is used in MA, or, to some degree how much impact it has. We don't have any "sorties flown" stats, so I had to come up with some kind of substitution. Counting kills only would result in exactly what you apparently like to avoid: An overestimation of planes like the 262, which happen to have very few deaths.
Planes like the F4UD are more numerous like their pure "kill" numbers indicate, because they are often used for ground attack duties.

But here you go, a chart which only accounts for kills:




And for comparison Kills&Deaths again:



Not that much difference at the top, but the 262 goes up by 16 positions.


And regarding to the 262's lethality... don't limit your view on pure fighter vs fighter combat. The 262 can go anywhere and strike anwhere at will. It can quickly go and hunt that goons... or try to take out that buffs that are about to level your field before any other interceptor can reach them.
Sure,  if you are in a fighter and aware of it's presence the 262 poses much less a danger than a Tempest...but the game is not all about dogfights

[Krusty]

Originally posted by Lusche
Counting kills only would result in exactly what you apparently like to avoid: An overestimation of planes like the 262, which happen to have very few deaths.

True, but when comparing against the P-51 kills, the P51D has (let's say) 30,000 kills, and the 262 has (let's say) 1000 kills.

The P-51D is 30x better. However, if you compare kills/deaths, the 262 is suddenly the second best fighter by a long shot. I'm thinking specifically of this chart:




Neither is perfect, as you've mentioned the flaws in the "kills only" ranking, but the same goes for bombers. You can't count how many times a bomber has flown around, hit nothing, and landed, or ditched, or just pulled the plug. That would skew those results a lot.

Neither is perfect, but I think "kills only" is a bit better for comparing the planes against each other, as in these charts.


EDIT: P.S. The charts are pretty cool, though

[Lusche]

Originally posted by Krusty
True, but when comparing against the P-51 kills, the P51D has (let's say) 30,000 kills, and the 262 has (let's say) 1000 kills.

The P-51D is 30x better. However, if you compare kills/deaths, the 262 is suddenly the second best fighter by a long shot. I'm thinking specifically of this chart:

Well, with the K/D chart I was just presenting MA K/D values. Not ranking planes in dubious terms of "quality". The "better fighter" discussion is a complete different topic. The K/D chart can only be used as one argument in it.

[Lusche]

Originally posted by Krusty
Neither is perfect, as you've mentioned the flaws in the "kills only" ranking, but the same goes for bombers. You can't count how many times a bomber has flown around, hit nothing, and landed, or ditched, or just pulled the plug. That would skew those results a lot.
 

This is true, though I sometimes have the impression that very few bombers actually return to base unharmed. (That 30k start raiders are rare species and one can argue about their impact on game)
Of course this isn't true for the Arado, which usually can easly avoid getting killed.

[hubsonfire]

I was being facetious- the 262 is a beast in terms of overall lethality. I  wouldn't suggest furballing in one, but with regards to simply getting kills, without getting killed, it's certainly one of the best. I'd agree that it's not the ideal plane for the average MA engagement, but that's another discussion entirely.

[dtango]

Nice work with the stats Lusche !

Food for thought for you:

[list=1]

• Using kills+deaths as a proxy for usage assumes that each kill and death are worth the same amount of usage.  Each death equals 1 usage point, and each kill is worth 1 usage point.  This means that 1 kill is equal to 1 death.  

The problem is that's not the case.  We know that for some planes you can get more than 1 kill per 1 death and in others less than 1 kill per 1 death.

One way you can adjust for this is to multiply the kills by the k/(d+1) ratio and then add the adjusted_kills+deaths together.  The k/d ratio tells us that 1 death = X number of kills.  Multiplying total # of kills by k/d ratio of that plane then normalizes kills so that you are counting 1 kill = 1 death.

EDIT: adding the below items:
=======================

• The stats show the compilation of the total numbers added up over 2007.  When drawing conclusions from this fantastic data we should remember that over the course of each tour those numbers are probably fluctuating as well.  I'd be curious about median usage and the deviations per tour per model.
• We should also remember that comparing %'s can be a little misleading because they mask the scale of the differences.  A 1% differential may not seem like a lot but if you have 10,000,000 kills+deaths in a year (not out of the range of possibility for AH) then 1% is 100,000 kills+deaths.
• It would be interesting to see aircraft series grouped together as well - for instance all spits as 1 group, all mustangs as another group, etc.

Again, fantastic work!

Tango, XO
412th FS Braunco Mustangs

[Krusty]

That's an interesting thought there... Trying to normalize kills. Not sure if that's the best way to do it, but interesting idea!

[Lusche]

Originally posted by dtango
Nice work with the stats Lusche !

Food for thought for you:

[list=1]

• Using kills+deaths as a proxy for usage assumes that each kill and death are worth the same amount of usage.  Each death equals 1 usage point, and each kill is worth 1 usage point.  This means that 1 kill is equal to 1 death.  

The problem is that's not the case.  We know that for some planes you can get more than 1 kill per 1 death and in others less than 1 kill per 1 death.

One way you can adjust for this is to multiply the kills by the k/(d+1) ratio and then add the adjusted_kills+deaths together.  The k/d ratio tells us that 1 death = X number of kills.  Multiplying total # of kills by k/d ratio of that plane then normalizes kills so that you are counting 1 kill = 1 death.
[/B]

That's more the direction of assessing some kind of a planes quality. I once dabbled a bit with it, but I didn't find it worthwile. I wasn't interested in that.

I rather see any kill as a single event. My "usage" table just states in how many events a certain plane had a part in - either as killer or victim. I did that because I can't quantify the number of sorties in any way.


EDIT:
OK.. I did it, just for fun:



Now, does this chart really tell us more? Looking at it one could think Tempest's have currently a huge impact on game, or you constantly run risk of getting shot down by them. Which isn't true at all. This charts does reflect actual numbers much less than (K+D) or kills only.

[Mr No Name]

Good topic and research Lusche... Thanks