• Update 0.31 is expected to rework more than half of the existing aircraft in the game. In order to update this guide and data with this, we will be taking our time to ensure the data is relevant post bugfixes and balancing.
  
  
• This might take anywhere from a couple of weeks to a month, depending on the scope of the balancing changes and the amount of time we can afford to allocate to this project.

• We were looking around for actual data on the flight performance of aircraft and couldn't find anything concrete, so we felt like we could contribute a little by setting some benchmarks.

What is the aim of all this

• To hopefully improve the awareness of everyone around the capabilities of each aircraft using quantified measures of performance that can be related and easily read by everyone.

Why is this formatted like this?

• While the guide is formatted similarly to that of an academic paper to allow for people in the future to easily verify any data or methods used in the making of this, we are also painfully aware of how utterly ♥♥♥♥ difficult it is to read those types of articles. This is a guide, not a paper, so this is made to be as easy to read as possible.

• Methodology and Raw Data contains all the details for people wanting to replicate, verify or otherwise inspect the raw plotted outputs from these tests.
  
• Limitations discussed our known issues with our testing methods and how that affects the resulting end plots.
  
• Data Highlights [Important!] is where pilots go to see the capabilities of each aircraft: Max Rates, Optimal Speeds and Turn Radius.
  
• Analysis [Important!] is where we discuss things we noted during our testing that you might find interesting or important.
  
• Future Works and End Notes is where we draw the line on what you can expect from this in the future and what will not be tested by us.

• There are 3 measurements taken at each velocity and altitude for each aircraft, with them being: Sustained Turn Rate, Instantaneous Turn Rate and the Turn Radius. The method for each will be discussed in detail separately.

Sustained Turn Rate

• Likely the most sought-after piece of data and the easiest to measure, this is measured by flying an aircraft using mouse control with zero centring force at the desired altitude.
  
  
  • The throttle is set to maximum, and the hardest sustainable turn (no losses in speed or altitude) is maintained. Stability Assist and Auto Vectoring are kept on for all aircraft.
    
    
  • For some speed, the control input can be maxed, and the speed still increases. In this situation, the throttle is decreased until a sustained turn is maintained.
    
    
  • A timer is set, and the time taken for the aircraft to perform a full 360-degree turn is recorded before being later converted via 360/time to produce the final Degrees/Second turn rate.

Instantaneous Turn Rate

• This is the most difficult and inconsistent on a per-run basis to measure due to the lack of a data logging mod or system for the game, so this is our best attempt.
  
  
  • An aircraft at the set velocity and altitude is flown in a straight line due north. At the start of a timer, the pilot attempts to turn as hard as possible due east, keeping the vector on the horizon as much as possible.
    
    
  • After 3 seconds, the change in bearing is noted, and Angle/3 gives our best guess at the instantaneous turn rate. We are aware of the limitations of this; however, please refer to the limitations for why we chose this method.

Instantaneous Turn Rate without Stability Assist

• By popular demand, this was added to test the full limits of each aircraft. This meets the same struggles as the regular test, except with even greater instability.
  
  
  • An aircraft at the set velocity and altitude is flown in a straight line due north. At the start of a timer, the pilot attempts to turn as hard as possible due east, keeping the vector on the horizon as much as possible.
    
    
  • After 3 seconds, the change in bearing is noted, and Angle/3 gives our best guess at the instantaneous turn rate.
    
    
  • This is then repeated 3 times and the average of the 3 tests is taken, with the maximum deviation from the mean set as the error bar, and the average being the final listed value.

Turn Radius

• This required two players to measure and was the most tedious of the bunch due to requiring multiplayer and fiddling with the map.
  
  
  • The same method as the Sustained Turn Radius is used to keep an aircraft in a continuous turn around a point (this was completed at the same time as the Sustained Turn Rate tests).
    
    
  • Using a custom map with radar stations spaced 100m apart horizontally on the grid near the flying area, a scale was produced for the map as well as a series of reference points.
    
    
  • A second player monitors the aircraft, and at the points where it is facing due south and due north, the location is noted, and the distance between these 2 points is measured and divided by 2 to give the turn radius of the aircraft.

• All data was recorded into a spreadsheet with the velocity, altitude, fuel load, and weapons load noted
  
  • Weapon loads were tested using the Default loadout as of [0.30.94], with the sole exception being Medusa, which was also tested with a full external jammer, Radome and internal ARAD loadout.
    
  
  
• Upon completion of testing, the data was downloaded and imported into Python using Panda and plotted using Matplotlib to produce the graphs shown further down.

---

---

---

---

---

---

---

An important note for future reference, as seen in the data here, the difference between 100m and 1000m is extremely small. This shows that the game has likely modelled its atmosphere to be similar to real life, so future testing can omit much of this and test only one of these altitudes (500m) to save time.

• Since diagnostic tools are not directly provided, the resolution of the data is limited by what we can reasonably measure consistently and repeatedly. These are the error margins for the data gathered.
  
  
  • Sustained Turn Rates are measured using a timer, and during repeated tests, the margins are within +-1 second (worst case). This comes out to a +-5% error margin, around +-1 Degree/Second
    
    
  • Instanaeous Turn Rates are also measured using a timer; however, due to reaction time and the small time period, the percentage difference between runs is much more significant. The data here is also at half resolution compared to the other results, with tests only at every 100km/h and results interpolated in between data points. This makes it the least reliable result on a micro scale, but since it forms part of a larger pattern and testing was consistent, on a macro level, this is still usable.
    
    
  • Turn radius is measured against a series of radar stations aligned 100m apart. The maximum resolution of the scale is 100m; however, it was possible to roughly discern if a measurement was between two stations closer to the halfway mark. Therefore, the results are accurate to the nearest 50m.

What does the Instantaneous turn rate show?

• Since there is no way to log data directly, it's hard to measure the true maximum turn rate. This is why we chose to measure it at 3 seconds, since that is the shortest time period we can consistently measure.
  
  
• The measurement of the turn rate at 3 seconds, we believe, is more relevant than the actual maximum turn rate, as it represents the "burst" turn rate of an aircraft during high stress, such as for missile evasion. This also represents the "snappiness" of an airframe, making it a useful measure regardless.

Are the testing conditions consistent?

• While we try our best to keep conditions consistent, time limitations and the lack of tools for this make that extremely difficult, particularly with how fuel-hungry aircraft like Revoker can be. To not go quite insane, we opted to set acceptable ranges for the variables tested.
  
  
  • Test Speeds are limited to +-25km/h outside of which a test is void
    
  • Test Altitude is limited to +-100m
    
  • Fuel levels are limited to +-10% of the testing levels, so tests start at 60% and end when 40% is reached.
  
  
• While we acknowledge this is not ideal, sadly, we are limited by our mortal bodies and minds, and these are the compromises made to allow this project to be finished.

Why are the low-speed results so weird?

• The lowest speed tested is not the lowest speed an aircraft can go; it is simply the lowest speed we can generally repeat.
  
  
• The proximity to stall speed causes difficulties in testing, and so the lowest speeds have much wider error margins, the amount unknown as we are often fighting to keep the aircraft in the air at those speeds so often the the "acceptable range" was often a suggestion at times.

For any questions, you can find me lurking in the Nuclear Option Discord as @thesharkdrowner, if I have the time, I will try to answer any questions.

---

---

---

Red = Stall + Crash after 3 seconds

---

---

Red = Stall + Crash after 3 seconds

---

---

---

And extra charts we plotted out of interest because we saw them somewhere before.

• Loadout weight matters FAR more than raw airframe performance
  
  
  • A Loaded Revoker will be out-rated by a clean Vortex despite turning 20% faster.
• Ifrit on pure theoretical performance is dominating with a huge lead over all other Turbofan aircraft
  
  
• Vortex flight performance is noticeably poor compared to aircraft in comparable roles, with Ifrit turning 26% faster than it.
  
  
• Medusa has a surprisingly high turn rate, with its performance primarily limited by its flight computer, which locks the turns at a certain G load.
  
  
• Revoker and Ifrit turn rates are pilot G load limited, as at high velocities the pilot will reach the 6.5G limit and be unable to sustain a full turn, meaning much of the aircraft's engine power remains untapped for sustained turns.
  
  
  • This contrasts with Vortex, which can maintain a full afterburner turn for all speeds and benefits greatly from doing so.
  
  
• While on paper the Deg/Sec difference between aircraft appears to be small, with only 1 or 2 degrees difference in some cases, it's important to note that in a paper dogfight with a 180 degree phase difference on merge a few degrees difference leads to significantly less time for the slower turning aircraft to respond before the other reaches a gun solution.
  
  
  
  
• This chart shows how long an aircraft has in theory, assuming both pilots are flying at their respective optimal speeds in a 2 circle and maintaining a sustained turn rate, before an Ifrit will catch up and gain a gun solution. This assumes a best case scenario for the defender with 180 degrees difference between the aircraft in a merge.

• Vortex and Revoker dominated in this field as they appear to benefit greatly at low speeds with Stability assist disabled.
  
  
• Ifrit performed poorly by comparison with little to no gain in performance with stability assist off. This is the result of the testing methodology requiring a 3 second turn, which due to the large surface area, Ifrit is incapable of performing at the speeds where stability assist is relevant. Noted in the area in red on the charts, Ifrit stalls its right wing during these turns early on, resulting in a flat spin before the timing ends, so only the maximum turn in that time is recorded.
  
  
  • I have no clue how reflective this is to people's experiences; however, this behaviour was consistent across the 3 tests. I am interested in hearing people's experiences with this as it is the only aircraft to see little to no benefits from stability assist being disabled.
  
  
• Aircraft do not universally turn better with stability assist off, with Compass notably performing worse on average at high speeds with stability assist off. This is likely due to porpoising and instabilities.

This project took over a month to complete, in between other projects that we are more obligated to complete. As noted from the data, future testing will see lower altitude resolution due to the ♥♥♥♥ all extremely minimal difference between results.

Here are the things we will almost certainly not be testing:

• More custom loadouts.
  
  • This is just the line we are drawing to prevent mission creep; the custom loadout for Medusa was an exception since the profile of the aircraft changes drastically with its addition.
• Testing without stability assist
  
  • Post-stall is a complex topic, and consistently testing and quantifying results is considered too time-consuming and is beyond our understanding of the topic, so we are abstaining from this topic. This has now been tested.
• Rotary Winged Aircraft
  
  • Oh hell no, I don't even know where to begin with this witchcraft

Here are the things we want to test someday:

• Higher altitudes
  
  • While not currently relevant in the meta (for dogfights), we think this might become useful at some point, so higher altitude results seem useful to have.
• 100% and 20% Fuel Tests
  
  • Lower resolution testing to see how much fuel mass affects flight performance is likely our next goal.

End notes:

Thank you, Shockfront Studio, for creating this game. We look forward to seeing and testing the creations coming out of future updates.

We will attempt to keep this guide updated for as long as possible; however, we are people and have other responsibilities, so there is a limit to this. Future results will be of lower altitude resolution as established, but more importantly, we draw the line on what we are willing to support at physics engine overhauls / complete game rebalances.
Retesting every aircraft?

Well, we don't really wanna to say the least...