Jump to content


  • Content Count

  • Joined

  • Last visited

Community Reputation

36 Excellent

About flagdjmetcher

  • Rank

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. I play SP only. In my first few hundred career mode missions I found the spotting very stressful and confusing. That seemed to match what I read about the experience of RL new pilots, and I appreciated the realism. In other words, I expected it to be very difficult, and it was, and I felt that to be an immersive experience. I'm aware that at the detail the spotting is not realistic, but the overall experience and level of difficulty I think is not far off. Certainly all the difficulties that I've had in missions are things that I've read were quite common IRL. After several hundred missions, spotting remains difficult for me. This is also realistic - with my eyesight and reflexes, I would have been pure cannon fodder IRL. However, it's no longer fun. I've had the confused new pilot experience, I appreciated it, but I've had enough of it. Fortunately as a SP player, I can turn icons on and off at will, and I nearly always do turn them on at some point during a mission. I think this gives me a reasonably realistic simulation of the vision and SA that a young, fit, 20/20 vision pilot would have. Slightly better for close contacts, and slightly worse for distant (given that icons disappear after 10km). So basically I appreciate having the options for both realism and fun. Better spotting would be nice, but even with that I'm probably still going to want icons on occasionally, and even with icons on I find this sim more enjoyable and immersive than The Other Sim With Icons. As for MP: A server full of fearless high-time pilots in immaculate late-war planes all doing their best to look for trouble is pure fantasyland, so the whole realism debate is a bit bemusing. However: if better spotting improves the MP fun factor, then I'm all for it, but I probably still won't participate.
  2. Maybe first and second derivates of all of the above as well? Sounds hard to me, but I'll defer to your and Patrick's greater experience. Also worth pointing out that it's not either/or. Even for humans there's a rules-based part, and then there's the bit where no expert can really explain why they did what they did. Both humans and state machines at some point have to rank their alternatives and choose one, and that weighting function can behave chaotically (in the mathematical sense) wrt win/lose or live/die, so maybe ML can help with that part of the problem. I don't claim to know where the devs could or should put their effort, it just seems to me that with so much data available from every sortie it might be worth throwing it at a model. Would be interesting to know what approach is being used for autonomous fighter research IRL. I guess we'll find out in about 100 years.
  3. I think ML could be useful. A state machine is fine, but knowing what state you're in is the hard part. A human can look at a flight of allies engaging a flight of enemies, and have a reasonable idea of who is engaging who and prioritize enemies based on level of threat to self and wingmates. That's really, really hard to do via a state machine, but quite a reasonable problem for ML. Once you've sorted out what's happening and where you can best intervene, I agree that the next steps can be a bit more deterministic.
  4. The MiG-3 in winter white with the red stripe is the epitome of fighter pilot cool. It also suits the way the AI flies better than the 109. Flying a 109 against AI Migs is noticeably more difficult than flying a MiG against AI 109s.
  5. I read an account recently (I think it was in a book on the Schweinfurt/Regensburg raid) about a Fw190 sneaking into a formation of escort fighters. The flight leader only realized what was up by counting the number of planes and finding he had an extra, who dove away as soon as he turned towards it. It must have been a fairly open formation, but still it's clear that ID at moderate ranges wasn't always straightforward.
  6. I just had a career mission in BoM in the bf110 E-2. It was an "Intercept ground attackers" mission. I was #4 in a flight of four. Difficulty mode was "easy". We approached the target area at 500m. When we got there, there were half a dozen IL-2's doing attack runs, and at least four MiG-3's flying escort. Might have been more escorts, but definitely had four at once in my sight picture at various times. I had to retry this mission seven times, and even then arrived home shot up and barely landable without having done much to interrupt the ground attack. I tried a bunch of different approaches: 1. Sticking to my flight leader. I chased MiGs off his tail a bunch of times. Unfortunately he didn't return the favour. I had a bit of success turning across the nose of other 110s and using them to scrape a MiG off my tail, but that was highly situational. I certainly couldn't master the art (if there is one) of staying tight with my Rotte while also maneuvring defensively. 2. Going hard after the IL-2's. Predictable outcome. Bounced by the escort every time. 3. Going hard after the MiG's. Pretty difficult with worse speed, climb and turn. I would have to be much, much better at long-range deflection shooting to pull this off. I did miraculously land some hits on the MiGs, but nothing that appeared to affect them at all. 4. Leaving my flight and going high before the target area, then attempting to BnZ. This was better. My high speed deflection shooting is so bad it was also completely ineffective, but that's just a skill issue. However, eventually a MiG decided to climb up and get me, and that was me done. So what to do? First up, I think we never should have engaged from low level. We should never even have approached the area from the level we did. Second, I think we shouldn't have engaged at all. We were clearly badly outmatched. The escort fighters were just flying rings around us. OK, this is war, Germany expects every man etc, so maybe we could have done at most one or two BnZ passes starting from up high to see if we could break up the IL-2's and bag a couple. But then we should have disengaged. Third of course would be nice if my flightmates would watch my six. But that's all AI behaviour beyond my control, and this isn't supposed to be an AI whinge thread. I actually think the AI is bloody amazing even while being somewhat terrible at times. I can't even imagine how hard it is to get something like that right. More importantly, what personally could I have done better? 1. Turn around, go home and fail the mission (can you get shot for cowardice in career mode?) 2. Stick to my solo BnZ plan, but bug out after a couple of passes? 3. Just keep repeating the mission until I survive by luck? 4. Do <I don't know what> better?
  7. Re subscription models: perceptions of acceptable pricing are pretty much detached from reality or consistency. I know people who were outraged by the Photoshop move to subscription - yet happily shell out for Netflix, Spotify, and their super-size-me phone plan. And then there's the people who complain about game prices in the $50-100 range and avidly wait for sales and specials - and then blow all of that and more on a weekly basis on booze and drugs and partying and uber-eats deliveries. When they're not buying gold and premium content on their favourite pay-to-win game. Which is their choice, but it doesn't fly well with the mock outrage at game pricing. Point is, I'd hate to be a game developer trying to navigate the total irrationality of consumer price perceptions. Speaking as a software developer, if you wanted me to develop something as big and complex as a new plane for you, my first priority would be to make sure you understand what you're asking for and what it's worth to you. My second priority is to establish whether you are a trustworthy entity. These are non-trivial tasks that I'm not getting paid for. The reason it's important is that if you decide at any point that you're aggrieved, you will eat my entire profit and more just in arguing about silly crap. It doesn't matter how "iron-clad" (lawyers will chuckle into their Armani sleeves at this) the contract is. You will still put me out of business just in the process of having my lawyers explain the contract to you - again. So, someone turning up at the door waving cash and begging me to take it is actually a business risk for me. For a small business, it's possibly an existential threat. It's one every business has to navigate every day, but we do so very, very carefully.
  8. As a rule of thumb, I find the mission briefing says we'll cruise at 375km/h, we actually cruise at closer to 350km/h, and I can sustain just above 400km/h with fairly relaxed engine settings and without anything overheating. If you're struggling to get to 350 something is wrong with your settings, and if your flight is streaking away at > 400 I don't know what's going on. As Mr Mobile BBQ says, control movements eat your speed. It could be something as simple as standing on the rudder pedal without realising (don't ask me how I know). I usually fly with 100% fuel, oil radiator closed and water radiator at 20-40%. I back off on the RPM and mixture to 80-90%. Anecdotally it seems that max RPM is bad for overheating, but I haven't done any real testing. I know when I get a 109-induced interest in how fast I'm going and crank the RPM and throttle up, I'm immediately opening radiators as well. Don't take this as best practice, just a reference point.
  9. Looking at the IL-2's record, they made the right call. The German tank crews seemed not too worried by the IL-2's; they could be relied upon to miss, and their own losses were horrific. Meanwhile in the West, the German mobile forces were all but paralyzed by the comms campaign and jabos with plain old bombs. By the time that Panther got to terrorizing the PBI in the village - well, it probably was just one Panther, most of its supply train was lying shot up in a ditch, and whether your rocket/bomb CAS can knock it out directly isn't really the point. What they can do is knock out the supporting arms that are preventing you from outflanking it. This is completely unlike the CAS doctrine from the Korean War onwards, where the idea is to offset a numerical disadvantage with the ability to very rapidly switch firepower from point to point. Hence firebases and gunships, and eventually A-10s.
  10. If you are the flight leader and there aren't too many fighters you can to some extent protect your flightmates in that situation by turning towards the incoming fighter. Kind of a Thach weave where your wingman has no idea what you're doing. Your flight will wobble all over the place trying to follow you, and once you've got enough IL-2's going in different directions it's easier to get angles on the fighters and the attacking AI seems to be less keen. Whether your flightmates will take cannon shots at it I'm not sure. If you're not the leader though you just end up alone at the back, which is not a healthy place to be..
  11. Nothing you say is incorrect, but you leave out the fact that per aircraft available and per aircraft lost - the 110 had at least as good a record as the 109. The loss rate per deployed aircraft was high - it was a valuable and heavily utilised asset - but the loss per kill was on a par with the 109 and Spitire. I'm not suggesting that, as an airframe, the 110 is a standout fighter. Post 1940 it aged badly. But in context, in that situation, it was a good fighter. The numbers don't support the argument that it performed dramatically worse than the 109. I'm just making bald statements here, but this is a forum post, not a scholarly article. The sources are there. Bergstrom is one of them, and _his_ sources check out, so I'm inclined to believe him.
  12. There are schools of thought. Its actual combat record in 1940 was good. As a fighter, its 1940 record is at least as good as the 109, as an escort it was better, and as a ground attacker it did very well. The only complaint the German commanders had about it was that they didn't have enough of them.
  13. I don't issue attack vectors, in absence of such I've noticed IL-2's make a concerted effort to make their attack run into the prevailing wind. Took me ages to figure out why they would fly past the target initially, and why they wouldn't do a bomb run straight down the road on convoy attacks, but that was it. Target facing doesn't seem to come into it; even with targets like dug in tanks where there's clearly a weak direction they just stick with upwind runsl. Maybe 10% of the time some of them will go a bit off-script, haven't worked out yet what causes that. Just my observations with undirected IL-2s, might not apply to other situations.
  14. A map grid should be achievable and not too unrealistic. I wouldn't expect a colleague in a whirling dogfight to be calling out map references, but when the whole flight has taken off somewhere one of them should be able to call out a rally point. And late war in the west, from what I understand, it was the norm to be vectored straight onto an ongoing fight by a radar-equipped ground controller. The interesting thing is that it's not consistent. Which I don't mind at all, seems realistic. I've certainly been bounced by AI that flew right up to my six without a peep from my in-game pilot, and (contrary to all the complaining about omniscient AI) I've returned the favour. I even managed to join an AI formation without arousing suspicion. The annoying part is when my in-game pilot is calling out threats from 10km all the way down to 1km and I don't at any stage manage to see them, but that's mostly frustration at my own bad spotting. The other side of this is that I've learned that the best way to find the enemy is to follow my flight leader. They see contacts waaaay before I do. Which is also realistic (given I'm a new pilot), but it does mean I essentially don't scan for threats at all anymore. Last mission reinforced this for me. I was in a flight of four heading home from an escort mission. My in-game pilot was calling out enemy fighters 10km to our six. Then six km, then five. Nobody in my flight reacted. Finally I couldn't stand it and did a 180 and went looking. My flight kept on straigth ahead. In-game pilot keeps calling out, all the way down to 1km. I see nothing. I turn around again and head for home. Just as I get to home plate, I hear my fight calling out "engaging the enemy". How on earth I got home before them I'll never know. I do another 180, backtrack, scanning high and low, nothing. Back and forth a couple of times, enjoying the radio drama. Finally I give up and land, and in the post mission event log I see that my flight had engaged the enemy just one grid to the west. If I'd looked in the right direction at the right height and at the right time, I would have seen them. So pretty much at every step in that whole process, the outcome was bad. Not really the game's fault, but at every stage I can think of a more useful game behaviour than what actually happened. So after that I've resolved to just ignore my virtual pilot, not bother trying to spot, and just follow my flight leader. Which is pretty much what rookie pilots were told to do, so - quite immersive I guess.
  15. Without wanting to open the PTO can of worms... "From 10,000 feet, scanning the airfield through my binoculars, I saw three aircraft - or dummies. Even with an excellent pair of glasses [RN for binocs], it wasn't always possible to distinguish between them". Norman Hanson, "Carrier Pilot", p264. In this quote he's flying a Corsair over Sakishima in early 1945.
  • Create New...