Vedexent

Maybe I shouldn't jinx it, but I don't see crashes more than once every 2-3 days, and that's usually when I'm trying to heavily use the system for other things on top of running the server. True, we're operating under modest player loads, and with missions specifically designed not to be "crash bait" - extensive use of triggers, etc.

Light bulb! Now this makes perfect sense. If DCS World is written in C++ - and I seem to recall that at least part of it it - then yes, there could definitely be memory leaks. Throwing more RAM at the problem would make the OS run better, but not the program. I suspect - as I'm a *nix guy and don't really know the internals of Windows - that you might be putting enough memory in to have some left over when all the allocated memory area for DCS world is "leaked away". DCS World would still be "broken", but it wouldn't take down Windows down with it. If DCS World is "leaky" it means that restarting, or rebooting, might be the only cure (short of the dev team plugging the leaks, of course). This would be true of Servers (probably more so) as well as people using the software to "just play". It may be that servers not only need to rotate missions every X hours, but do a full system reboot - or risk stability issues. Not ideal. I wonder if one could throw a system monitor in there with the logic: "If no one has been on the server for X minutes and the server has been up for Y minutes OR the server has been up for Z minutes, regardless of current activity, reboot Windows" - and have DCS World auto setup as a server? But ... can you setup DCS World to automatically connect as an MP server and launch a mission on startup?

At the risk of stating a "less filling / tastes great" kind of debate, I like the aircraft of FC3, and play it for that reason - so I still think FC3 has value. I would, however, be thrilled if the Su-25(SM?) came out in full DCS world fidelity. I won't argue against the idea that the DCS World level modules are "better", by my standards at least. Still, I'm not expecting a click-able super high fidelity Su-25 anytime soon. Barring that, I'm waiting on the MiG-21.

And let's remember the adage: "be careful what you ask for, you might get it!" I'm sure that enough money, resources, developers, and time would be thrown into DCS World to fix all the issues and make is a fully stable, bug free platform if EA bought ED :P Sometimes the cure is worse than the disease.

If the system is having next to no problems in single player, and it's having no stability problems in multiplayer when you are the server (you might experiment with that if you have not), then it might be a networking issue. What kind of connection, speed (up and down), latency, and reliability is your connection? How many machines are sharing that connection in your home, and what are they doing? Do you have anything streaming video like YouTube or NetFlix? Do you have a VOIP phone? Do you run a torrent program (for open source software images, and public domain material, of course). I have a new, decent, but not spectacular gaming system (i5 4670, 8Gb RAM, Gigabyte Nvidia GTX geoforce 670 2 Gb of RAM, and a pretty normal commodity non-SSD hard drive in my system) but I have a very good network connection for my area (50Mbps down/ 10Mps up, unlimited traffic, very consistent throughput, excellent latency), and I have zero problems with Multiplayer on other machines, and I've only had problems being the server when I've tried to do 20 other things while other people are using my server and switching back and forth to DCS World. It's true that traffic is modest on our server, usually only a handful of people and we've peaked at a dozen. I keep missions rotating reloading every 3 hours, and I'll typically restart the whole server first thing in the morning when no one is flying. It could be that your system is fine and network capacity, consistency, or saturation issues, are clobbering your connections.

So - initial question solved! ;)

I agree 100%. I won't hold my breath ... but a full fledged Su-25 or even the Su-25SM in full DCS World level fidelity would be great. Maybe by 2035 :D

Welcome:) Bottom line: what kind of aircraft appeal to you? You don't really have to buy anything if you're willing to fly the Su-25T, and don't need to get on the Multiplayer servers. If you want to get on the Multiplayer servers, you'll have to have at least one module activated - I think, someone correct me if this isn't true? - but you can get the A-10A and the Su-25 for about $USD15, right now on the ED store. So - you can play for very little up front investment if you like. If you're willing to sink more money into it, you can get any of the high-end, high fidelity modules like the Mi-8, A-10C, Huey, or Black Shark. It all really depends on what kind of air combat roles appeal to you. Want to fly ground assault? A-10A, Su-25, Su-25T, and A-10C Want to fly Helos? Huey, Mi-8, Black Shark Want to fly fighter aircraft? Look into FC3, but be aware that these are simpler flight models with simpler cockpits. Want to fly a WWII vintage fighter? Mustang Also, leave budget room for hardware: you'll at least need a joystick like the Logitech Extreme 3D (sort of a minimal entry level, IMO). You can spend up to $600 for the Warthog Hotas setup, although it's very nice.

Makes sense, and explains why my current practice with AAA of breaking left 45 and pulling up hard, works - it's kind an odd blend of these two tactics. But it makes more sense to be doing your maneuvering/acceleration in the plane perpendicular to his line of sight like you describe, instead of half-in, half-out like I'm doing currently. I'm curious why you say "Dive not too steep", though - unless it's a speed/stability issue. I've been doing high, sharp dives to stay as high as possible, as long as possible, and to fire from above. I'm finding that long shallow dives take me within range of all sorts of ground units, and bleed off the speed.

I agree 100% with the idea of saving speed and/or altitude, I've only been braking to stabilize the targeting reticule during rocket attacks. I can definitely see the advantage of egressing quickly out of the engagement envelope, but I'm wary about getting bracketed against the ground if SAM AAD units open up; hard to drag an IR SAM down and back up when you're 30m off the ground. I'm assuming that people are inverting to get greater seperation between their plane and the flares? Typically, when I'm climbing out of a rocket or gun run, I'm pulling an Immelman - trying to convert that speed back into altitude and folding my course back over my tail, then a quick aileron roll and I'm set for the next run, pointing (roughly) at the target zone - and from an F3 camera view the flares are going quite wide of the flight path.

Distilling all the advice I've gotten, both on and off the thread, that's now pretty much what I do for air-to-ground rocket attacks: high altitude, high angle attack dives, back on the throttle - sometimes even popping speed brakes to keep the airframe from starting to shudder while I'm lining up the targeting reticule and awaiting the "launch authorization" LED to light up - periodic scans for white SAM plumes, then fire, pull in brakes, slam the throttle to max, pull up and 45degrees away from where I think the SAM threat is, popping flares in tight groups of three. If spot a launch, it's hard dive (brakes in, throttle up), popping flare triplets, roll so the SAM is on my beam, wait for the rocket plume to go out, then pull up perpendicular to the path of the SAM, high throttle, and stand her on her tail, flaring all the time ... and pray. Seems to work well. I've also started to have some success with the aforementioned tactic of running an Avenger out of missiles and them clobbering them; started taking a pair of Kh-25Mls along with the S-8KOM pod loadout for stand off attacks on Avengers/Chaparrals. I don't know how long it takes vehicle based AAD units to resupply. Radar guided SAMs, I should at least get a lock/launch warning.

Just one more data point on the "I can't leave DCS" front: on multiplayer it seems to matter whether or not the server is paused. Paused = can switch to browser window; Running = sometimes can't switch.

Something else we've been playing with - although it's dangerous and shouldn't actually work on an intelligent crew: run an Avenger dry. Taunt it on the edge of its range, make it fire at us, evade, repeat. When it's fired all 8, zoom in and kill it. Clearly a human gunnery crew wouldn't likely fall for this. It's much more dangerous than the two plane "you bait it and find it; I'll kill it with a Karen" approach, but a single plane can do this ... or get blown up trying.

Not part of the detection discussion, but relating back to the OP, it also appears that the automated ptogram for dispensing of flares is woefully inadequate. Timing flare until the missile is close, and then popping off a rapid cluster of 3-5 seems to be much more effective.

Have you checked it out in the camera display view? If you have the pro-clip, or the reflectors, set in a way that when you move your head all the way down, one of the points eclipses one of the others, then the TrackIR can wildly misread your head position. You need to view the camera view display, then move you head through the range of motion side to side and up and down, to make sure this doesn't happen. You may need to reposition the pro-clip or reflectors.

Ah, playback errors. I remember the first time I rewatched a mission and went "wait a minute! I didn't blow up there!" :P ----- It seems that the TacView program "hook" exports and saves game data pretty much like people seem to expect a "saved game" function to work. I don't know if it stores enough variables, or whether all the required variables are exposed to grab. I also don't know if it would be possible to re-create units with enough fidelity by using scripting functions and libraries like Mist. Still, it looks like a third party plugin is implementing a large part of what people seem to want in a save game feature: if the system design won't do something you need, hack around it :P

Yes it does - and I'll have to pay attention to the track IR issue as well. I know that if I'm just leaving the "server" running, going to spectator mode, and then going off to do something else, it seems to be less likely to "lock" if I F10 to the map view in the background of the server screen before alt-tabbing out; that could be related to TrackIR.

OK - as an addendum, I'm able to replicate the OP issue. I upped my graphics settings, set DSCW to full screen (hadn't realized it was in windowed mode because it was also set to the full screen resolution) and enabled vsync. I've now gone from not being able to switch away some of the time (under windowed) to not being able to switch to some of the time (under full screen). Either way, there's a problem here. Doubly so as I'm using this system as a personal computer, while keeping a DCSW "server" running on it 24/7 right now. So far as I can tell it's still running, but I can't use it :p System: Windows 7, i5, 8Gb RAM, Gigabyte Nvidia Geoforce GTX 670 card Edit: It also doesn't seem to matter if I invoke multiplayer from the single player menu, or directly start multiplayer. Edit: Of the two "issues", the one that seems to be the least disruptive is windowed mode, full screen resolution, vsync on. There are times when I can't get out of DCSW without shutting down the program, but this seems to happen much less than the OP issue under full screen.

True. I guess I should have said no dedicated ground assault aircraft like the Su-25. Everyone's going multirole these days - even the A10C is facing phase-out.

I also pulled images of some RCAF modelling decal sheets that might be helpful with signage ------- With all this in mind, I think there are three basic livery strategies here. Green/Gray cammo Multi-shade "ghost" low contract cammo Blue-on-blue "naval" cammo To that end, I've taken a stab at pulling some color pallets from the diagrams. 2-color green and grey (front line green) RGB: 68,87,59 - #44573B RGB: 95,99,110 - #5F636E 2-color green and grey (front line green: lighter) RGB: 68,87,59 - #44573B RGB: 165,163,164 - #A5A3A4 3-color Ghost camo (low contrast) RGB: 133,131,132 - #858384 RGB: 165,163,164 - #A5A3A4 RGB: 134,143,152 - #868F98 4-color blue and grey (naval) RGB: 47,63,97 - #2F3F61 RGB: 61,84,118 - #3D5476 RGB: 107,125,149 - #6B7D95 RGB: 144,160,175 - #90A0AF I'm not a graphic artist, by any means, so those color palletes may need some "tweaking". If I were to pick just one pallet, I'd go with the last one - maybe selecting either the top or bottom three colors in a 3-color cammo pattern. I really appreciate anything anyone does with this :)

Wow - that's a very generous offer to at least try. I'm familiar with your skin work, I've got some of it installed on my Su-25s - especially the historical Russian skins. :) ---- Basing everything off the CF-5, there were three main squadrons that used that aircraft 433rd Tactical Fighter Squadron (porcupines) 434th Operational Squadron (bluenose) 419th Tactical Fighter Training Squadron (moose) Squadron patches were as follows: CF-5 Liveries appear to have been like this: Actualy "field aircraft are as follows" Bluenose Moose Porcupine?

This is s what I'm coming to believe as well: any MP mission that strives to have any decent level of "realism" pretty much needs to be multi-role co-op play. Maybe it would be different if the AI was more reliable, and you could set up a full AI-on-AI military action, where AI units would drop out as player units took over their roles (and presumably re-enter if the player drops out). That kind of relegates open servers with "drop in and play, then drop out" types of missions to not very realistic mission scenarios, or "practice range" scenarios. Not to say those types of scenarios can't be fun; just that they give up realism for playability. High realism scenarios seem limited to per-arranged co-op events.

I'll give it a try when I'm back home again. I'd like to see if there is any effect changing weather conditions around, as well. I'm by no means saying the detection model is perfect, or that under certain circumstances might not be exactly that "super alert". I just know that it seemed their detection was worse in a thunderstorm.

The coding project sounds ambitious - and very welcome if it adds realism to the game. But, are you sure the detection model is that simple? Given the exact same map, with Chaparrals and Avengers in the region, the same area seems very much less dangerous to fly in a thunderstorm than it does in a clear summer afternoon; much fewer Stinger launches. Now "seems" isn't data, and I wouldn't bet the mortgage payment on that being the case, but I would have thought the units were having problems visually locating the nearby attacking Su-25s. Edit: That's not to say that your improved detection code isn't needed! I'd love to see it and test it out.

VapoR and I did some experimental flying on the 314th server tonight, on this front. Doing the pair hunting, with one plane baiting the AAD unit SAMs at the edge of their engagement envelope - and then flying defensively and escaping - and the other plane smacking the AAD unit down with the Kh-25Mls seems to work decently in game (provided the "bait" pilot is experienced enough to have a decent chance of surviving the SAM encounter) - although I don't know if that's terribly realistic. IRL I would think this would work only with inexperienced AAD unit crews; veterans would "sucker you in" closer before clobbering you. Auger: I really like your point about needing to be way more familiar with the target zone and expected enemy forces than the typical "I load up my plane with every munition under the sun and run out to the waypoint to see what's there" kind of MP approach. I'm finding that with the practice range missions we have on the 314th, we've flown some of them enough that we have familiarity with the target zones to that level, and they are actually survivable now, because we have that kind of familiarity. It might be interesting to pick a "green field" target zone - known to the mission designer but not the pilots/ground commander - and use a combination of terrain study, and CA unit recon, to see if we can gain that kind of familiarity before ever putting planes in the sky. Then do a "single pass" attack, called in and guided by the ground commander - JTAC style - and then pulling back again. With decently experienced pilots who have a good change of killing their units on a single pass, I wonder what kind of results we'd see.