Piloting Tip Needed: The Last Of The Lightnings

[fallout3 9]

    After several years, I have again taken up the miraculous mission pack: The Last Of The Lightnings, made by Comrpnt. As some of the missions require formation flying and pairs landing with AI controlled Lightning, I feel the need to ask for piloting tips from you tigers, as I now know that I am a pilot of ham fists as I always suspect 

    The main observations after 5 sorties for the first mission in the pack, a 2-ship circuit, and 1 sortie for the last mission, the 9-ship flying pass, turned out the following questions and observations, which might be useful for our discussion of the extrodinary mission pack, and some aspects of SF2 more generally:

1. Unusually Fast Accleration Uring Take-off Roll

    The AI acclerates very fast during and immediately after take off. First thing I do after mission loading is shoving the throttle to Max and release brakes. My Lightning starts moving when the AI lead is still static (so I actually develop full engine thrust before him), and at the inital stage of takeoff roll I could close the distance and separation to some extent. However, the AI lead could still dash away afterwards, and I am not able to stay quite close later in the take-off roll.

    I suspected that this has something to do with friction of the runway slowing my Lightning down, as there is no way to confirm the take-off technique the AI uses, and possibly he has applied backpressure quite early to lift his nosewheel just off ground, so to reduce the total friction on the wheels, even as this might introduce more induced drag. Thinking about this, I tried to ease nosewheel just off ground earlier than I could observe any sign of rotation of the lead Lightning, there is no telling if this works, as there is no way t precisely time the events during the whole take-off phase, but the AI lead could still consistently out-acclerate me despite my engines reaching full reheat earlier than him.

    Seen from the rare in this screenshot, the lead seems to have his main wheels a little off ground? Or it`s just an illustion caused by view angel and the Lightning`s contours? I`m not so sure of anything now.

2. Very Fast Acclearation Away Without Reheat After Take-Off

    Immediately after lift-off, the AI climbs very sharply, and it is not possible for me to keep very close to him for fear of acclerated stall. During the last phase of the take-off run, when lead is already in air and my aircraft is just above to clear ground, he`s almost right over my head instead of above and ahead of me. I understand that this is probably the result of him curving his flight path at linear speed not so much higher than my own, or even lower, as he gets rid of the ground friction but gets at the same time much more induced drag with his sharp pull-up. I could understand that AI does not give so much consideration to his human wingman, as if he does he should be climbing-out straight and shallower for wingman to catch up and adjust formation. 

    However, the AI will make a sharp turn right after this, with reheat off, and with reheat on I could not keep up with him acclerating away until a while after, so the distance and separation of the formation will enlarge quite much after take off, due to the lead`s quick, acclerating turnaway that needs not even reheat.

    This is the most I can get out from the pairs takeoff, and it is very busy. He gets aways from me without needing the burner plume at all.

3. Suspected Unusual Decent and Deceleration Ability During Turns

    During turns in the circuit, the AI seems to be in the habit of acclerating rather quickly at the beginning of the turn, and is apt to decend at the same time, and at the end of the turn, they seem to have the habit of idling the throttle. I usually fly on wing outside of the turn, and am of the impression that during a turn I should fly just a little higher and faster than the lead, so that we will turn at the same angular speed, but I at slightly greater linear speed due to the greater turn radius. This is quite hard to actualize, though, and as soon as the turn begins, the lead seems to be able to make it a descending turn, and at the sametime very tight, so if I try to match his angular speed, the lift component in the vertical will be stop my machine from decending in the turn with his, and the lateral separation increases as a result.

    The other issue is that the AI seems to like idling, or at least retarding very sharply the throttle at the end of a turn, and it would be very easy to overshoot. However, in the final 9-ship fly-past mission, it is not hard to stay close in formation during the shallow, gentle turns, so I need to confirm these turn-related observations by starting the mission and staying on the ground, while using F6 to observe how the AI solo the circuit all the way, lest they are only visual illusions caused by perspective during a tight formation turn at large bank angle.

    This is result of the 9-ship flypast. I could manage the turns better than in the pairs circuit, but certainly hope to be able to turn in as close formation as when flying straight and level, for more RAF style.

4. Tremendous Deceleration and Abrupt Pitch Change Before Approach

    Right when the AI lead reaches the approach waypoint, he will decelerate sharply, and with some abrupt pitch and altitude changes, without even deploying the speedbrakes. In a welded-wing close formation overshooting will be certain, as retarding the throttle quickly to idle cannot match such great deceleration, and the speed brakes will cause quite much pitch change. At the same time, as the new speed is quite low, flaps will be needed, or the aircraft will start losing stability. The solution I found is to loose the formation somewhat before the anticipated lead deceleration, and use speedbrake right away when seeing the lead "stops" in the air, and lower gear and flaps at first chance.

    The speed brake will be kept out for rest of the approach and landing, as this seems to be the only way to adjust formation during the pairs landing, by leaving you the possibility of using throttle. Without the boards out, there seems to be every chance for you to creep ahead of the lead. Pairs landing is possible and consistent in this way at satisfactory formation distance and separation, but still needs care when it certainly should in SF or not.

    This happens when I sense the AI lead is doing his dread deceleration.

 

    If I manage to keep formation after the magic slow-down, a formation like this could be managed right to touch-down.

    My thought is that certain AI behaviour is coded rather than aerodynamic, especially in accleration and deccleration, and during take-off and approach. This may be a trade-off to make AI able to fly everything. It seems to me that AI is flying every landing like carrier landing, keeping just above the touchdown speed, at a pitch angle and thus AOA that does not allow for seeing the runway sometimes, and without control and throttle corrections, along a precise glideslope. This technique is very hard for players to follow, expecially someone like me who likes too (probably overdemanding) fly in formation with AI take-off to Touch-down. 

    By far the best game in this respect is MSFS, and at a certain update for IL-2 1946, 4.09 or 4.10? I could manage to land in pairs with even an AI I-16. However, now in IL-2 all that an AI can do is to make a strange, probably hard-coded turn seemingly along an invisible axle attached to the trailing edge of rudder into "approach mode", and let down all the way at precisely the touch down speed, nose-high, to an stalled 3-point landing. SF2 fairs good enough as I could pair land with an Lightning, or F-5C with some struggles against the dreaded "approach mode" deceleration, but we could certainly use more probably.

    This experience with Lightnings also led to some questions. We know that there are values given for critical speeds like landing or MaxSL or take-off or stall, but how does they affect AI behaviour, and what other possible factors and setting affect AI behaviour, is still in question, at least for me in my ignorance. If only a take-off speed is given, but not landing speed, then how will AI decide its own landing speed. Is the take-off speed for rotation, lift-off, or climb-out? Similarly, is landing speed approach speed or touch down speed, and what should the value be for aircrafts that are supposed to be at different speeds for crosswind and final leg? Is the stall speed for all-up or gear and flaps down, power on or power off?

    Additional questions could be asked on how the AI respects these numbers or not, and if it does, to what extent. Will AI force itself not to exceed the maximum speed SL even if it could? Will it do its best to keep above the dictated stall speed? And the most curious question among all is how many AI behaviour are hard-coded and to what extent we could change that. It would certainly be nice if we could work around all we could by certain clever tricks in number tweaking and mission editing.

    I guess it is already a lot of questions and a lot of words from my part indeed. Time for some screenshots of my hairy attempt at the legendary Lightning, and I will be most thankful for your thoughts and comments on the potential issues, not necessarily covered above. As I felt that something might be due to my clumsy formation turning techniques, any piloting tips or your experience in flying with the AI will be more than welcome. 

[+yakarov79 10,758]

I think you shoot the most serious questions in 2023...and in recent  years at all...

Everyone is speachless...(?)