I have now;
--
Nate,
Re DCS:-World Low-level flight.
Initial observations:
I made a test mission very similar to the previous one I used in FC2.1 and DCS:-A10C; A flight of 2 x Tornado, then 2x F/A-18C, then 3x F-15E - all following the same waypoint path.
The low-level flight profile for tactical fighter/attack aircraft now more-or-less works within DCS: WORLD. It's is a significant improvement to tactical flight planning. However this now casts the limelight on the low-level problems that were also present in DCS:A10C, but were absent from FC2.1.
Namely, the problem caused by open-formations in deep-close valleys.
But before I go into that;
First, I'm going to make a suggestion about the new low-level mission planning waypoint editing tab:
(1)
The "Waypoint Settings" area of the mission planner is now much less buggy and glitchy, with better low-altitude setting options. I particularly like that the altitude reset 'bug' due to terrain height effects is now gone when you move an existing waypoint around the map (thank f**k for that!). And also seamlessly inserts the last-used setting in the next waypoint. This is much better.
Now it mostly works like it ALWAYS should have.
(2)
Throttle control and engine spooling by Ai is much smoother, thus the use of afterburner is much reduced and improved in DCS:-WORLD. FINALLY! So now there's less fuel being unnecessarily wasted by Ai flights.
However, the Ai still use airbrake much too often, bleeding energy for no clear reason, then 'requiring' more fuel to re-accelerate. This likewise should be addressed because the Ai aircraft at times repeatedly open airbrakes for a second or so, then immediately close them again, and again, and ... it's quite dippy - please quash this.
Air brake is for when you REALLY need to slow for essential reasons, like tanking, landing, ground attack approaches, reducing speed to reduce turn radius, to increase instantaneous pointing and agility.
In other cases the Ai aircraft should 'coast', or, if appropriate, climb slowly to wash off unwanted speed, as altitude can be turned back into speed without using more fuel. The realistic flight-model and flight envelope is not worth much if the Ai 'flyer' is not flying realistically and efficiently in the way a human would.
If they do, then the real performance capability of the jet, it's effective achievable tactical range and available tactical loiter, will have effect in missions - as it should.
More efficient flight and fuel use also means the fight can use more flanking and high-speed manoeuvre, and achieve more altitude, as well as orderly RTB, to the correct planned base landing point.
Plus they are then less prone to suddenly bug-out, due to early bingo, abandoning the planned return waypoint route, that the flight's planner designed to reduce vulnerability and provide some protection so that they wouldn't be targeted during RTB. Which was a real problem with all previous versions.
So any improvement to fuel use by things like sensible use of airbrake and of coasting will have effects on tactical outcome plus help broaden-out tactical planning possibilities and reliability of the Ai.
The more the Ai flight is optimised for realistic flight behaviours, the better.
(3)
The actual flight at tree-top-level is now very good, EXCEPT for one important aspect. Aircraft approach ridgelines very well. They rise up its face properly, they seem to avoid structures over towns, and then they fly over the top of a ridge at genuine low-level, with flat aspect and profile - all very good. BUT! ... the flight then tends to remain above the terrain that's now falling away rapidly beneath it, so can still end up a few thousand feet AGL, for 20 seconds or so.
It seems to be less of a problem over very steep terrain, where they tend to swoop down more, but more of a problem over flatter coastal terrain (where you actually do need them to get down low fairly fast).
What the Ai should do is to more quickly commit to a "swoop" manoeuvre, down to a level significantly below the ridge-line, rather than taking their time to get down there again after the ridge has been passed.
More swoop please!
It's vital the opponent's EW radar and AEW struggle to paint the flight. Lingering nearly level with the ridgeline is not the way to achieve radar-masking, and to frustrate successful engagement by SAMs.
Overall low-level flight gets a score of 7.5 / 10 from me, which is up from a dreadful 0 / 10 within DCS:A-10C
I say only 7.5 because there's another vitally important issue that needs to be rectified within DCS: WORLD, one that was not an issue at all in FC2.1, but that will have to be fixed.
(4)
Regarding this, I'm going to suggest an important new inclusion is needed within the "waypoint settings" area, within the mission planner.
First, I need to describe the problem that's now occurring in DCS:-WORLD namely;
Although the low-level flight basically works the first mission I made quickly highlighted that a low-level flight using waypoints down the very centre of a narrow twisting deep valley, is quite incompatible with the standard open-formation flight in DCS:WORLD, even when using just two aircraft within a flight.
My last test round, a few months back, within DCS A-10C, also demonstrated that these lose open-formations are NOT VIABLE for low-level flights of more than one aircraft.
And the more aircraft in the flight the more acute the problems become, because aircraft #3 and #4 end up flying kilometers away from the plotted waypoint, and up along the TOP of the ridge-line (!).
Thus completely defeating the purpose of low-level tactics--to deny the enemy information about where you are, what you are doing, and how to kill you. So yes, the low-level flight is fixed, but the formation flight issue has yet to be fixed to the point where you can actually use low-level mission planning of multi-ship flights.
The solution is obvious;
The mission planner needs a NEW checkbox option within the "waypoint settings" area, so that the mission builder can select "line a-stern formation" (or alternatively, "close formation") as an actual option to be selected between planned waypoints, such as when flying down a valley. Then de-selecting it (i.e. going back to a standard lose open-formation) when executing the final run-in on the assigned target(s), etc. Thus you need "Line A-Stern" formation up to the IP for low level flights, then open-formation for the final lead in to the attack (ingress), then the next waypoint must converge on a "line a-stern" formation again, as egress occurs, until the flight is clear of threat zones, and steep terrain. Then to use open-formation again once the flight has exited high-terrain areas, and for flying circuits to a landing, etc.
"Line A-Stern" formation is necessary as it is in fact the one that's commonly used by military fast-jet flights, in deep and narrow valleys, as these U-Tube videos all depict:
http://www.youtube.c...e&v=fPgvld0rr0AThe bottom line is, no one tries to fly a low-level line-a-breast open-formation through narrow valleys. There are scores of videos of air force's practicing low-level flight through mountain regions, and it's all Line-A-Stern, with flights of two or more aircraft typically displaced by about 500m behind the lead aircraft (probably to get some smoother air).
Plus there are several compelling 'wingman-stability' issues demanding a line-astern formation option for low-level flight waypoints.
Nate, I'll describe next what commonly occurred during low-level tests within this new DCS:-World, by breaking it down into separate digestible elements:
(a.)
As in DCS:A-10C the flight struggles to remain 'cohesive', as an open-formation, because the guy in the valley floor (following the exact waypoint path plot) must fly slower than the guy on the side of the ridge.
This seems to be because the turns in the valley are sharper for the guy down in the valley floor, so the LEAD uses the waypoints and the terrain map to anticipate what turn radius and speed will be needed to achieve the upcoming turn, at the next waypoint.
As a result it appears the Ai in DCS:-WORLD is looking at the next waypoint turn and is independently (and actually correctly) judging and adjusting 'his' speed to a level lower than that which was set in planning, in order that the turn can be made, within the available space to do so, at that altitude. The tighter the turn and the closer the waypoints the slower the LEAD Ai pilot in the floor of the valley must fly the turn.
This is actually pretty clever.
I like that it now does this, but it also becomes problematic as for one thing, it means that if you insert a cruise speed of 420 knots into your flight plan, for the entire flight, but this Ai LEAD dude instead thinks it can only manage 330 knots (on average) in practice, then the mission-planner's time estimates of when you'll reach your IP, ends up being several minutes too optimistic.
So you don't get there when the planner says you will, in fact, you're going to be very late.
That makes it much more difficult to plan any highly coordinated multi-flight and multi-type as well as multi-role attack and 'supported' and covered strike, from different directions, with down-to-the -second planned TOT, for each flight, and each member of each flight, for maximum assault impact (and also for de-confliction).
The solutions are obvious though;
(1) The planner must use a few less waypoints (so not track the valley floor's path so accurately) and to leave more room to get the turns done in particularly tight valleys, by cutting the corners a bit (which unfortunately increases the chance of being detected and targeted) and/or/else;
(2) To reduce the planned transit speed for the whole leg to the IP, so that the time of arrival estimates for each waypoint become more or less accurate;
All though this can still potentially generate problems for the Ai; namely, if the flight is then to slow, and it is cutting the corner of a spur, and it is hot a day, and the weapon and fuel load is near MTOW (heavy), then the available lift is going to be much lower, and the manoeuvre room and options in tight terrain become even more reduced and constrained.
And when you do this, the flight of the aircraft does indeed become increasingly less stable and erratic, as the speed decreases to less than 300 knots.
Again; the answers are available and logical; you must therefore carry less heavy weapon loads (so you must be more precise and adept with targeting and optimising the launch conditions to get a hit), and you must aslo carry less fuel, so plan to use a tactical tanker on RTB (and try to fly only on cold days ... I keeed!).
So the planner can probably allow for all these time and speed factors in planning the flight and coordinating the attack phase. It is of course never going to be perfect in the real-world either, but real-world digital autopilots do in fact deliver split-second TOT capabilities, so the developer should also attempt to provide that to the tactical SIM-er.
I like it though, as it increases the practical and mental challenge of creating a low-level attack plan that will come together and work, plus het you back on the deck in one piece (I just hope I can master refuelling techniques again). It will definitely give a thrill if you get it all 'just-right'.
This ability is what I really missed in LOMAC and FC, as it just didn't work (but in earlier strike combat SIMs from other producers it did all work), so this is getting a lot closer to working properly, within DCS:-World.
It's closer, but its not quite there, yet.
(b.)
But what occurs with the open-formation, is that the 2nd wingman tends to always fly faster than the LEAD aircraft, down in the valley floor. Or more accurately, on average the guy who is higher up the valley wall tends to fly faster than the guy lower down, within the valley.
So #2 may fly 2.5 miles past the next waypoint BEFORE the LEAD even gets to it, and #3 will go even further! So what then happens is, #2 does an afterburning MID-LEVEL U-TURN, at high speed (i.e. large radius, so flys higher and higher, often above the ridgeline!) then comes screaming back along the roof of the valley (i.e. exposed to AEW, EW, SAM & fighter sensors who know are alerted that something is going on down in that valley) then #2 does another U-turn once it gets back to the waypoint that it missed going through, well behind the LEAD.
This is FUBAR Nate.
But it gets even worse because #2 then usually accelerates to catch up to LEAD, but then doesn't slow down, so rockets forward past LEAD again, at >600 kts. And the further the flight goes down the valley in this open-formation, the worse this gets, because then #2 finally burns up enough fuel from A/B use, that if begins to fly slower. But now it loses the energy and lift and thrust needed to fly OVER the spurs, that are jutting out from the ridgelines ...
KA-BOOM! ... crashes into ridge! ... dumkoff!!! ... lol
So the open-formation in low-level flight ... HAS TO GO!
It will need to be replaced with a waypoint-planning section option for an open "Line A-Stern" formation, BUT, where the flight's overall speed is constrained and governed by the LEAD aircraft's speed (as we can't have #2 becoming a butt-plug for #1).
Obviously the developer's DCS programmers have attempted to allow individual aircraft, within each flight, to remain lose and somewhat autonomous, whilst in transit, and to not necessarily stay line-abreast, nor even to do the same speed. i.e. they have tried to remove the 'fake' looking scripted moves that forced wingmen to act as definite in-sync formation with respect to the LEAD aircraft.
The result of this is an indefinite and unpredictable flight formation, with regard to position and speed.
So what happens if you forced that somewhat randomised speed behaviour into a "Line A-Stern" formation? Will #2 then stay behind the LEAD aircraft?
I don't know, but I suggest the developer looks into this very soon, because this is what must happen, in order to get the low-level flight planning and tactics sorted out properly, so it will work with flights of more than one aircraft. Which is obviously essential.
So we will need a "Line A-Stern" formation option as a tick-box, to be integrated into the Mission Planner's "Waypoint Management" tab to get this working in planning, and in actual flight. As those videos show, low-level line-a-stern open-formations are tactical de rigueur for any air force flying down narrow valleys.
(d)
Getting back to that #2 plane, that crashed into the mountain side ... the reason why that happened is quite simple, the Ai tried to fly over it when it was in no condition to make it, it was too heavy, and too low energy and slow, plus not enough thrust to weight available to push it over the top.
BUT, the real problem is this; it didn't TURN AWAY from excessively steep terrain while in that energy-incapable flight condition.
It just kept on going ... like the little train the could ... it was so sad. </sniff>
Just make them turn away when in the state the developer. Make them try a different path.
That's what a human would do. The solution is that simple.
But of course, this situation wouldn't have even developed if #2 had not been charging up and down the valley with A/B lit, due to these open-formation issues.
(e)
My final point, for now, is related to something I saw in the Su-25T Campaign, namely;
Su-25T at about 1,500feet AGL fires at SAM site.
SAM site naturally fires back at Su-25T.
Su-25T pulls it into steep missile-defeating fast dive.
But Su-25T completely fails to stop diving toward ground ... KA - BLEWIE!
No more Su-25T!
My suggested Ai missile evasion response tactic would go something like this:
(1)
Tell Ai that below 1,000ft AGL dives must not exceed -30 degrees.
Tell Ai that below 500ft AGL dives must not exceed -15 degrees.
Tell Ai that below 250ft AGL dives must not exceed -10 degrees.
etc.
(2)
Tell Ai that below 250 feet to pull-out and offset 135 degrees to the SAM and increase to max power while making vigorous and rapid yo-yo vertical manoeuvres, while flying along that 135 degree offset to the SAM launchers, between a max of 500 feet AGL, and minimum of 100 feet AGL, while popping chaff and flares, as necessary.
(3)
Tell Ai pilot to head for the nearest negative-aspect terrain, within three miles of its forward 60 degree FOV, once in this 135 degree offset to the SAM launcher.
(4)
Tell the other Ai wingman in the flight to lend support and immediately attack the SAM site's radar trackers.
--
I also made a few missions with SAMs and SEAD aircraft and found that many of the SEAD jets were crashing into terrain or even into water much too often when trying to dodge Buks and Hawks.
They don't all crash, as some do manage to pull out of the dive in time, but a lot do crash. This missile evasion manoeuvre (script) is going to have to be adjusted so the aircraft pull up a bit earlier.
I'll do some experiments with planning multi-type low-level strike/attack packages as these may actually work ok, IN MORE OPEN TERRAIN, where there are no steep or narrow confines that interfere with open formation transits to target.
I won't bother testing the low-level flights down deep valleys any further, as it's thoroughly dysfunctional, until a line a-stern formation is available for such flight plans.
I notice the SAMs seem to have a much slower reaction time, which is more realistic, and also good news if using a low-level approaches, thus making successful engagement more difficult, thus more survivable. But the SAMs do seem more lethal if they do fire at you and you are positioned in a clear line of sight (like when flying into a targeting and attack mode flight profile).
I think I also noticed that a Buk missile did not lose track when an aircraft went behind terrain. It appeared that it was still illuminated by a networked enemy EW radar (but not by a tracker), from another direction ... I think that missile should have lost track on the aircraft, but it continued to LEAD navigate the fighter, even when the fighter had dropped behind a ridge (from the F6 key missile view perspective).
If I see this again I'll make a track of it, as I forgot to save the track from it this time around.
Also, the fighter missiles, even older types, seem more lethal. Which is all the more reason to get the low-level radar-shadow use working and optimised properly for lights, to avoid detection and interceptions.
For some reason the F-15C seemed to be getting it's butt handed to it in an odd way, as it seemed reluctant to engage as early as it used to. Perhaps it's the response and targeting time also being delayed for it?
Dunno, I may be wrong about this, but the Eagle didn't seem as aggressive and bad-arse to me, as in FC2.1. Perhaps they are also looking and waiting for a better nav and kinetic vector for the missile engagement now? Whatever, but old Mig-23s with older generation missiles had no problem taking down the Eagles.
The AEW's racetrack loiter option is also a good one, it took while to find it in the planner though.
I can see the planner has much more specificity built into its planned actions, triggers and options now. Good, as long as the interface is intuitive, and bug-free (which I'm reserving judgement on until I know more).
It's going to take awhile to learn the new system, and what it can do, but it's nice to see it's much less buggy, and what I've tried messing around with has eventually worked ... so far ... but I haven't even looked at a manual for the planner yet.
The dynamic weather rocks, and it looks especially life-like. It's really well done. I was shocked by the new cloud formations and the whispy 'showery' appearance under some of the clouds in the mountain valleys and passes--love that!
I also like the smoke plumes from the burning vehicles I just think the fire and smoke disappears much too soon after a strike or a plane crash. I also think the new explosion smoke and dust fades away too quickly. It would be nice to have an option to vary the time such things are visible for. I'd personally leave vehicles and crash sites burning for an hour.
What I see so far in DCS WORLD feels much more compelling than in DCS:A10C and Black Shark, SIMs that I weren't that impressed with as a user and didn't find that interesting, so I stuck to FC2.1. Going by user polls many other users felt the same ambivalence toward them, but DCS:WORLD is much more interesting, for me, (i.e. not just another tired old re-hash of LOMAC ... Mutton dressed-up as Lamb).
I'll let you know if I find something else that is worth reporting, but please get the open-formation and controlled flight into terrain issues looked at, soonest.
zzz
