AI Aircraft Unrealistically High Energy Retention

[Civetta_ITA]

Hi all...
I'm writing about an issue that has been known for many years but still affects DCS: the energy management of AI-controlled aircraft.

I created a test mission: F-14 (human controlled) vs. SU-27 (AI set to Ace). If you look at the Tacview file that I've attached, the final part of the engagement is particularly telling, after the initial rolling scissors phase.

The SU-27 manages to follow the F-14 in the vertical despite starting with an initial energy deficit of 0.2 Mach (F-14 at 0.99 Mach vs. SU-27 at 0.8 Mach). When the SU-27 starts its climb, it's evident that it experiences virtually no energy loss.

Can anything be done to address this apparent bug in the AI's flight model/logic?

Thank you.

Tacview-20251003-141802-DCS.zip.acmi

[Civetta_ITA]

UP

[Gierasimov]

On 10/3/2025 at 2:35 PM, Civetta_ITA said:

Hi all...
I'm writing about an issue that has been known for many years but still affects DCS: the energy management of AI-controlled aircraft.

I created a test mission: F-14 (human controlled) vs. SU-27 (AI set to Ace). If you look at the Tacview file that I've attached, the final part of the engagement is particularly telling, after the initial rolling scissors phase.

The SU-27 manages to follow the F-14 in the vertical despite starting with an initial energy deficit of 0.2 Mach (F-14 at 0.99 Mach vs. SU-27 at 0.8 Mach). When the SU-27 starts its climb, it's evident that it experiences virtually no energy loss.

Can anything be done to address this apparent bug in the AI's flight model/logic?

Thank you.

Tacview-20251003-141802-DCS.zip.acmi 185.52 kB · 9 downloads

Right. Acmi file won't cut it as ED wants only track files. 

I often fight against AI so I know exactly what you are talking about, it is largely related to this feature that is in development (supposedly): 

https://forum.dcs.world/topic/373845-general-flight-model-development-progressany-news/