As requested, a central linkdump to all of the tests I've been doing with the aircraft of Warthunder. Here you can find relatively accurate performance stats for the aircraft in game, as they perform using Mouse Aim in Realistic Battles.
Any plane with a
strikethrough is an outdated test, performed before any recent changes to it's flight model.
He 112 A-0
He 112 B-0
Hurricane Mk II
Ju 88 A-4
Spitfire Mk IIa
Beaufighter Mk VIc
Me 410 A-1/U2
Spitfire Mk Vb
Fw 190 D-12
Ta 152 H-1
Spitfire F Mk XVI
He 162 A-2
Me 163 B
Outdated tests have been hidden under this thing. As you can see, many FMs have been updated since this whole thing began 6 months ago.
So, why spend all the time to make these?
There are literally hundreds of threads on this forum loaded with thousands of postulations and assumptions regarding the performance of various aircraft, but not that much documentation. Sure, the historical performance of these planes is pretty easy to google up, but after just a few games, it's apparent to many the planes in game do not behave historically, and the in-game stat cards can be notoriously inaccurate.
I've played this game for a few hundred hours and yet, other than a vague idea, even I didn't know how these planes perform either, and seeing thread after thread with nothing but assumption, prejudice and hearsay, this plane UFO that plane OP, I found I'd rather know the truth of the matter than guess. I don't really consider myself very good at this game either, but I found after just a few performance tests that finding facts is a bit more fulfilling than getting shot down. So, might as well get started and get some good data out there. For science.
Now, onto the methodology of these tests.
Every test is done with Mouse Aim, Realistic Battle flight engine, unlimited fuel and ammo and Auto Engine controls. All units are in Metric and speeds are True Air Speed unless labled otherwise. 30min fuel options are chosen where possible, to provide more consistency.
All models, unless specified otherwise, are fully upgraded and using Current settings.
, and not the Reference settings. The reason for this is personal experience. Not long after the 1.35 patch, a number of tests using the Bf 109 E-1 reference model showed it significantly underperforming vs the E-3, which should not be, since historically the E-1 and E-3 were extremely similar. In a QnA held shortly after the patch, the devs, in quite clear terms, stated that the E-1 was performing correctly. A short while later, a guy named Mattressi mentioned how one of his tests of the Spitfire Mk I demonstrated the reference model did not use the 100 octane fuel. Logically, then, other reference models might be missing upgrades. Sure enough, the fully upgraded model of E-1 performed nearly the same as the E-3, showing that the reference model's error was to blame.
Because of the potential of these errors, it's simply safer to use a fully upgraded model, a version representing the model players will actually be using in game, regardless of how historically accurate the setup might be.
Single Turn time tests are fairly straightforward, I aim the plane at a particular direction, at high alt I use the view cone on the minimap as reference and at low alt I use the red vehicle, slow down to 300 kmh IAS flying level and steady and I start a stopwatch. At 4 seconds i rev up to WEP and at 5 seconds I begin a level, clockwise, 360 degree turn, taking care to not gain or lose too much alt. At the end of the 360 when the plane hits the reference mark I glance over at the stopwatch and whatever is shown -5 seconds is recorded. If I have doubts of the accuracy, like if I gained or lost a lot of alt from start to finish, then I do it again.
Sustained Turn Time tests are similar. Start a stopwatch and do rings, keeping altitude level. With Mouse Aim it's just go in circles, note how many circles have been completed, and after 10 or so, note the time and divide by 10 and note the average speed at which the plane completed the turns. This gives the turn time that the plane can keep up indefinitely. This test is also done with Full Real controls, which I've finally learned a few things about, and involves the same procedure.
For climb tests, I slow down to the intended climbing speed, the default being 250 kmh IAS at 100m. I start a stopwatch, rev up the engine to WEP and climb at whatever angle necessary, with constant corrections, to maintain the intended climbing speed. Every 500m the time is noted. This continues for a maximum of 20 minutes. Previously, the climb rate was just calculated and graphed, but now the climb rate is normalized with a second set of climb data taken on a second climb test using the climb rate data shown on the browser map. It makes for much better looking and more accurate climb rate charts. The climb data is then extrapolated to determine the aircraft's Service Ceiling, sometimes listed on the charts as Max Alt.
Redline is found just by diving and finding the speed at which the speedometer turns red, signaling violent buffeting and potential breakup. Karaya_one tested the mechanics of breakup speeds a long time ago:
Breakup is the number listed as the "speed limit" when the plane breaks up due to overspeed, but as just noted, is well above the possible breakup speed. Tolerance is the maximum number of Gs that can be sustained when turning without the wings ripping off. This is usually found just by making a vertical dive and pulling up, with multiple tests to be absolutely sure of the correct number of sustainable Gs. WEP time is the duration WEP can be run until overheat messages appear. You can probably squeeze a bit more time in most cases but some won't want to take the risk.
The level speed tests are probably the most nuanced. To travel as level as possible at a sampling height, I use the Virtual Cockpit view and the directional indicator, zoomed in as far as possible. With Mouse Aim fully zoomed in, and the reference of the Virtual Cockpit's artificial horizon, the game reveals itself as having "notches" of direction in which the Mouse Aim reticle can fit.
These notches limit ones ability to pitch the plane. All aircraft wings generate more lift at higher speeds, so with some aircraft, a level nose may produce climb, instead of level flight. One notch will produce slight climb, while another notch will produce slight descent, so, at times, perfectly level flight at full throttle is impossible. In some cases, the descending notch will increase speed, and lift, enough to enable nearly level flight, but this isn't always the case. Where absolutely level flight is impossible, the speed at the notch which produces the smallest rate of change is used.
An example: If a plane, at 4500m sampling altitude, travels 554 at one notch with a slight descent of 1m per 6 seconds, and travels 550 at another notch with a slight ascent of 1m per 3 seconds, then 554 will be listed as the speed at that altitude, as it results in the smallest rate of change, and the most level flight. By luck, or chance or just having done dozens of hours I've found a way around the notches involving rolling the plane and just moving the reticle left or right, thereby enabling a full range of pitch. All measurements for speed are now made in while flying as level as reasonably possible at the altitude measured.
These tests are done with with the intent of finding facts, how these aircraft perform in game, not to find which plane is OP or predicting the next FOTM (though i'm sure the data will be used to such an effect). As the past year has shown us, change takes a long time, and the flight models we have today might very well be the ones we still fly a year from now. Better to find out how they fly now, sooner than later.
I sincerely encourage you to replicate or repudiate these test results, find any flaws or errors, refine the methodology and obtain more precise data through the use of other testing methods. This whole thing is about finding facts, so if you do a test and it's better than mine, or you find something wrong with my tests, then let me know.
Other planes will be added to the list as I do them. Each test takes about
5 7 8-10 hours from first takeoff to posting, the majority of that spent accelerating to max level speed. I've been averaging at least 1 a day but that won't last for too long. and can probably keep this up for a couple more months. I'm doing them at a leisurely pace since stuff in the internet spaceship game has been taking my time.
As of 30 Oct 2013 I don't intend to add more data to the charts. I know level acceleration, dive tests, corner speed and such have all been requested but at some point it becomes so much information that a test would end up taking 8 hours! My first tests of planes only recorded just Speed at Height every 1000m, G-tolerance, and Redline, so the charts have evolved a bit. I'd like to think there's a balance struck in the current incarnation.
What you see now in the most recent tests is the final format. In the end, I found it worthwhile to add all this stuff, which ironically means the tests take about 8 hours each.
If you have questions just post them here and I'll check in every now and then to answer. Also, if you have a request for a plane to test, post that too
and if I have it fully upgraded, or it's low enough level that I can get it upgraded in a couple games, then it will be put on the list. If it's not fully upgraded then just wait for it, I'll have it eventually. As of right now I have most planes unlocked and more than enough lions to just suicide into ground units all day for XP.
Alternatively, if you wish to join in with the testing, contact me via forum PM and we can work something out.
Feel free to check out Nabutso's Turn Time Chart
. His results are different than mine because of differing methodology, but the results are a fair bit more precise. Also laid out in a far more convenient fashion.
Special thanks to Slotter
for doing bomber and dive bomber charts. Truely, he is a hero for tackling these. And to Popingheads
for doing the Ta 152.