[Thoughts] DCS 2024: State of the Game

The previous article raised issues such as the AI of the ground and airborne units, radar fidelity, EW, AI AWACS, datalink, and so on really affected the overall feeling and experience of flying and operating a combat aircraft.

A little more than one year has passed, and I thought it was a good moment to see what has changed since then, analysing some of the points I raised at end the of 2022 and comparing them to what we have or what has been announced. This is, of course, not a comprehensive list. Before getting there, let’s open a parenthesis to discuss the ultimate effect of the mentioned problematics.

“The” Issue

My biggest gripe with DCS is its inability to make you feel part of a combat sortie. Bold statement, given the fact that I have never been in a military aircraft, so where does this conclusion come from? Rather simply, from discussions with actual crews, books, interviews, real footage, radio recordings, et cetera. There is, in fact, a common denominator that appears in most of the mentioned sources: uncertainty.

Here are a few examples to support my point:

• During the Vietnam War, the BVR capabilities of the F-4 and the AIM-7 were severely limited by the inability to assess the side of a contact with certainty. Specific “traps” and kill zones had to be used to ensure no friendlies were mistakenly engaged.
• The Iran-Iraqi war lasted for the majority of the 80s. When the Iraqis fielded the first Mirage F1, their RWR could not detect an incoming AIM-54 Phoenix. According to Tom Cooper, the Iranians were more concerned with the MiG-21 at that point in the war.
• Fast forward a decade, an F/A-18 Hornet was splashed by an Iraqi MiG-25 during Operation Desert Storm. Accounts vary, but it seems a case of ROE and delays in the fire authorisation.
• During the war in Kosovo, which started in 1998, several times NATO aircraft engaged the battered and outdated MiG-29s used by the Serbian Air Force. In one engagement, the AWACS had radar information about a possible Serbian aircraft but was unable to provide a reliable identification and classified them as “bogeys”. The F-15 later identified them as MiG-29s and engaged them.
  In another case, the AWACS was initially unable to spot the contacts. The F-15s had instead positive radar contact, and identified and then engaged them. In both cases, the F-15s used a combination of NCTR, non-friendly IFF interrogation and other criteria to escalate the contacts to hostile.
• In 1999, a US Navy F-14 Tomcats patrolling the no-fly zone in Iraq, launched an AIM-54 Phoenix targeting a lone MiG-23. The Flogger reversed its course and turned cold, whilst the missile eventually depleted its energy and crashed. Iraqi pilots often tried to bait and trap coalition aircraft.
• In December 2003, an American Patriot battery located near the Iraq-Kuwait border killed a British Tornado GR4 and its crew. Apparently, their IFF was not operating correctly, and the Patriot operators did not thoroughly vet their target. Erroneous identification was a tangible issue, and, in fact, another Patriot battery downed an American F/A-18 Hornet a few weeks later.

These examples are just a drop in the ocean. Still, they are very different and showcase events caused by ROE, “Fog of war”, lack of Situational awareness, target behaviour, avionics failures, incorrect settings, or simply inability to deal with the situation. Of course, no one wants to kill friendly aircraft by mistake, but the point is that DCS does not even allow such a remote possibility (at least in theory, let’s pretend casual servers do not exist for a minute!). Ergo, in the present game, a sortie can be messy because of the number of aircraft or the complexity of tasking, but unless a solid work of scripting and editing is done, it is rarely due to the difficulties of operating a military fighter jet in a hostile environment full of unknowns.

2022 vs 2023

Now that the general picture is clarified let’s discuss the points that still need to be improved to reach the goal defined by Eagle Dynamics itself: becoming the best combat simulation.

EW

Let’s start with the big one: EW, the acronym for Electronic Warfare. In DCS, the EW simulation is very limited. Understandable, given the delicate nature of the topic. What is less understandable is the current implementation. For instance, the emitters in DCS are limited to noise jammers. This is, in my opinion, perfectly fine. Next, the burn-through ranges are very static and set in stone. Quite a conspicuous simplification, but we can see the logic behind this decision. What seems completely arbitrary and difficult to agree with instead is the decision to implement jammers as 360° jamming strobes. I can see the coding difficulty of implementing directional emitters, as a fixed cone would make jammers pointless against SAM and non-hot, non-co-alt hostiles at low ATA. However, what we have makes very little sense: in real life, replicating what we have in the game is no minor feat. In fact, I doubt that real-life dedicated EW aircraft can match the jamming output of a DCS MiG-21.

AWACS, Datalink and Classification

According to the NATO website, the E-3A’s radar performs a complete rotation in circa 10 seconds. Therefore, the information it can provide and feed to datalink systems has a pretty significant delay. A contact would have all time to appear and disappear, using terrain masking, for example, or completely changing the geometry, thus heavily affecting the SA of the fighter. In DCS, AWACS updates information in real-time, fundamentally turning the SA page into an F10 map plastered in an MFD. This problem should be easy to tackle, perhaps with a “realism” checkbox in the mission editor options. Datalink, LINK4 and LINK16 are, in fact, other significant issues of the game, especially the latter. It is so unrealistically precise and reliable that certain modules can place the use of their radar, comms, and coordination in the background. The absurd effect is that many players are de facto addicted to LINK16 to the extent that the probably best overall air-to-air module, the Strike Eagle, is downplayed because it does not support LINK16 yet.
A more complex problem to address is the classification of a contact. DCS lacks native IFF implementation. Although some 3rd party devs have done some work in this regard, I can’t help but wonder whether this is not poorly used time: don’t get me wrong, I totally appreciate their effort, but shouldn’t a system capable of working cross-modules be introduced by Eagle Dynamics?
The lack of a classification system and the all-seeing AI makes actions such as Visual Identification and related intercepts pointless. In DCS, the classification is fundamentally bistable and denies an incredible amount of gameplay variations and depth.

RWR & Radars

Two more critical points are the radar simulation and the RWR simulation. In both cases, realism-wise, there is a vast discrepancy between different modules. The Hornet and the Fighting Falcon are still the worst offenders, but ED has taken concrete steps to improve the situation.

AI

The already mentioned all-seeing, simplistic AI is an old story. It sadly killed my enjoyment of single-player campaigns and missions. The ground AI is the least problematic. It mostly suffers from exceptional precision, awareness and pathing issues.
Air-to-air battles are, unfortunately, the worst offenders, as no smart tactics are performed by the aircraft, and increasing their skill level simply means that the AI cheats more. The best the AI can do is set up CC, gate, shoot, crank and dive. They offer no preemptive kinematical defence, there are no parameters to, for example, set up proper CAP tracks, grinders and so on. There is no difference between their tactics when they know or do not know the threat they face. Basically, all they do is charge in with a couple of variations. Any different behaviour must be scripted or addressed by the mission designer. Ah, and did I mention already that the AI disregards the effects of jamming? For instance, they can launch lofting AIM-120 beyond the burn-through range!

The other side of the coin

The discussion, so far, may have given the impression that the situation is quite poor. But is it, really? Actually, no. Not at all. The year 2023 has seen several improvements and peaks in features that should be close to delivery:

• The radar simulation is slowly improving, and the F/A-18C Hornet and the F-16C Fighting Falcon have received a few improvements in this regard. This, however, leads to the question of whether these updates will be applied to Flaming Cliffs 3 modules or not. If not, this would widen the quality gap between the DCS and the FC3 environments even more.
• One of the big questions I raised last year was about reinventing the wheel. Apparently, ED is making a huge and welcomed step to address this possible issue by adopting Heatblur’s impressive RWR simulation technology. Well done, ED!
  Possibly related, the SPO-10 RWR mounted in the Mi-24P is being drastically remodelled, and radar bands are slowly being applied to anti-radiation weapons as well.
• The issues with AI, starting from the ground, are being addressed. As part of the Dynamic Campaign improvements, ground units will receive new path-finding criteria, behaviour, formations and much more. As far as we know, the air side is also being improved.
• It is worth mentioning that ED is responsible for the core game technology and the modules it produces. Given the fact that they are not a huge corporation the size of Microsoft, delays in the creation of features and improvements such as Multithreading understandably take a conspicuous amount of time.
• Linked to the previous point, the impressive work done on the radio simulation and the comms between players is worth mentioning. SRS is still the best option, thanks to its flexibility and integration with other software, but ED’s solution is progressing very well.

Conclusions

All in all, it has been a positive year. The list of improvements I would like to see is extensive, but with Multithreading almost tackled, there is room for software and hardware-wise to add more and more of them into the simulation. We have not seen a lot in terms of core gameplay and mechanics. For instance, the Dynamic campaign was recently mentioned in a newsletter, but it is something I would like to see only when the issues mentioned above are addressed.

I look forward to seeing what next year’s “State of the game” will look like!