I agree with Mikaela and some others in most cases but someone may fail when it comes to explain what he/she feels and why.
Another guy (i don't remember his name) mentioned that you cannot compare 2 games in terms of reality, only real life experience to a game. I agree and from this point of view, Richard Burns Rally is the closest thing i felt compared to my on and off track experience.
Like in any real scenarios, when in doubt, follow the money, which in this case translates to the data that supports the opinion or "the feel".
Since CMRae in 1998, codemasters has continued to disappoint me with every new release, up to the current title. What Mikaela described as "submerged in liquid", some others have described it as "greased brick" with a central pivoted axis. I couldn't agree more.
It seems that is relaying on pre scripted tables which are made to match very rough numbers and don't interact with each other.
It is very hard to find the braking points or the racing line because there is no consistency with the "physics" and no feedback to and from the calculations (software). This is very obvious when cornering at low to medium speeds or power sliding, suddenly the car reacts abnormally and inconsistently with the input because of the current pre-scripted table switching and taking over.
A very interesting video which explains a lot about the sim-arcade difference (over 1 hour), Brian Beckman and the Physics in Games:
Some info about slip angle vs slip ratio:
http://www.edy.es/dev/2011/12/facts-and-myths-on-the-pacejka-curves/
This is what some of th
e other sims are using:
http://gaming.wikia.com/wiki/Comparison_of_racing_simulators
Nothing can be found here except the phrase "prototyping a handling model":
https://en.wikipedia.org/wiki/Dirt_Rally
Also nothing specific here except "The Ego engine was developed to render more detailed damage and physics as well as render large-scale environments":
https://en.wikipedia.org/wiki/EGO_(game_engine)
There is no physics specific information, the only physics showcase i found was about the destruction of obstacles in a race course.
Is the tire modeling (if there is any) Pacejka based or Pacejka-like simplified (no mainstream computer can run the full code in real time, 1 frame every few seconds) or something else and new?
This is why the developers of LiveForSpeed insist on simplified graphics for very high frame rate, anything less than 60Hz doesn't support sufficient calculations for physics resolution. Their new tire model is in development for years.
The technology behind Richard Burns Rally is also very advanced, even though it was half finished. If you look closely to the code, you 'll find that there are settings and switches that never implemented because of the deadline SCI forced on the team.
As of Dirt Rally, i think this is the best they can do, there are no physics to correct-adjust, it is the refinement of the old neo/ego engine (they call it "handling model ") which actually is a container of different modules for sound, graphics and... apparently physics! I would give my cat's right arm for a new rally sim, but as the old adage says, you can't polish a t**d.