Car performance accuracy report
-------
Car performance accuracy report-------
F1lrclub-vcr physics claimed to feature realistic car performance for grip, aerodynamics, power, and, as a consequence, realistic corner speed, straight speed and lap times.
As we promised, here is a statistic to check this mod's performance accuracy. Once the season is over all championships running F1lrclub-vcr ended, so we collected the relevant info available and compared to F1 real data.
Lap times and top speeds:
Notes:
-For the success of the analysis we use data from different Rfactor leagues (5), and also real Grand Prix data from 2011 and 2012 seasons in order to make estimations when 2013 data is misleading due to wet weather conditions or other.
For example, we don't use 2013 Melbourne real qualifying times because it was run under wet and cold conditions. We use 2011 and 2012 Q3 times to make a estimation of normal dry conditions time instead.
Source for real 2013 performance estimations: 2011 and 2012 F1 seasons.
Estimations for Rfactor when Q3 not valid: server records and Q2 times.
-F1 lap times are from official F1 site. F1 top speeds are mainly from telemetry data shown on TV (speed traps are not always placed at the fastest place).
-Rfactor data was available at liveracers.com or league's official results, replays and statistics. Most of information might be still available.
-All GPs held on 2013 tracks are included. Gps not included were not run, were run at different track layouts than 2013 (f.e. Singapur), or there is no data available.
-'X' means the value is considered 'correct', within the error margin (less than 0.5 secs for lap times, less than 5kph for top speeds)
-Race lap times are not checked, there was no championship running 100% race distance with the original mod.
( . . . . .Track . . . . . Q3 Lap time . . .gap . . .Q3 Top speed. . . gap )
-Melbourne
Real F1:. . . (estimation) Q3: 1.24.5 . . . . . . . Top Speed: 312-315
F1simrace:. . . . . . . . . .Q3: 1.25.6 . . +1 . . . Top Speed: 311 . . . X
SpainRacingFactor:. . . . Q3: 1.25.9 . . +1.3 . .Top Speed: 308 . . . -5
LigaRfactor:. . . . . . . . .Q3: 1.25.5 . . +1 . . . Top Speed: 315 . . . X
F1Latinoamerica:. . . . . Q3: 1.25.3 . .+0.8 . . Top Speed: 316 . . . X
-Malasia
F1:.(est)Q3:1.34.6 . . . . . . TS: 310-314
F1SR:. .Q3: 1.34.0 -0.5 . . . TS: 312 . . . X
SRF:. . .Q3: 1.34.1 -0.5. . . TS: 307-13 . X
LRf:. . . Q3: 1.32.9 -1.8. . . TS: 315 . . . X
F1L:. . .Q3: 1.33.2 -1.5 . . . TS: 313-17 . X
-China
F1: . . Q3: 1.34.5 . . . . . . . . . TS: 318-320
F1SR: Q3: 1.35.3 . . . +0.8 . . .TS: 320. . . . . X
SRF: . Q3: 1.35.7(est) +1 . . . .TS: 317-324 . X
F1L: . Q3: 1.35.1 . . . +0.5 . . .TS: 320-25 . . +5
-Bahrein
F1:. . Q3: 1.32.3 . . . . . . . . TS: 310-314
F1SR: Q3: 1.30.7 . . . .-1.5 . TS: 321 . . . . +10
SRF: . Q3: 1.31.0 . . . .-1.3 . TS: 313-318 . . X
LRf:. . Q3: 1.31.5 . . . .-0.8 . TS: 325 . . . . +10
F1L:. .Q3: 1.30.3 . . . .-2 . . .TS: 320 . . . . +5
-Barcelona
F1: . . Q3: 1.20.7 . . . .TS: 314-18
F1SR: Q3: 1.19.6 -1 . . TS: 322 . . . .+5
SRF: . Q3: 1.19.5 -1 . . TS: 318 . . . . X
LRf:. . Q3: 1.19.1 -1.5 . TS: 320 . . . .X
F1L: . Q3: 1.19.4 -1.3 . TS: 322-20 . .X
-Monaco
F1: . .Q3: 1.13.8 . . . . TS: 283-288
F1SR: Q3: 1.14.2 X . . .TS: 276-285 . X
SRF:. Q3: 1.13.9 X . . . TS: 282 . . . . X
LRf: . Q3: 1.14.5 +0.5 .TS: 288 . . . . X
F1L: . Q3: 1.14.7 +1 . . TS: 278-284 -5
Test: Q3: 1.13.2 -0.5 . TS: 280 . . . -5
Rfactor>>
https://www.youtube.com/watch?v=PDR_rRbrDGQ
Real >>
https://www.youtube.com/watch?v=DameemcGkgs
-Montreal
F1: . .Q3: 1.13.2 (est) . . . . . . TS: 322
F1SR: Q3: 1.15.0 . . . +1.8 . . .TS: 312-18 . . . -5
SRF:. Q3: 1.13.1 . . . X . . . . .TS: 315-16 . . . -5
F1L:. Q3: 1.15.1-4 . . . +2 . . . .TS: 321 . . . . . X
-Silverstone
F1:. . Q3: 1.29.6 . . . . TS: 308-12
F1SR: Q3: 1.33.7 +4 . . TS: 320 . . . +10
SRF:. Q3: 1.31.4 +2 . . TS: 315-18 . .+5
LRf: . Q3: 1.32.2 +3 . . TS: 314 . . . . X
F1L:. Q3: 1.31.4 +2 . . TS: 315-18 . +5
-Nurburgring
F1:. . .Q3: 1.29.4 . . . . TS: 300-302
F1SR: Q3: 1.30.4 . .+1 . TS: 308 +5
-Hungaroring
F1: . .Q3: 1.19.3 . . . . . TS: 308
F1SR: Q3: 1.19.5 X . . . .TS: 305 X
SRF:. Q3: 1.19.2 X . . . .TS: 308 X
LRf: . Q3: 1.19.8 +0.5 . . TS: 308 X
F1L:. Q3: 1.19.6 (Est) X . TS: 305 (Est) X
-Spa
F1: (est)Q3: 1:46.7 . . . . . . . . . TS: 323
F1SR:. . Q3: 1.47.1 . . . . . X . . . TS: 327 . . . . . . X
SRF: . . .Q3: 1:47.0 (est) . X . . . TS: 327 (est) . . .X
F1L: . . .Q3: 1:47.9 . . . . +1 . . . TS: 321-336 . . .+5
-Monza
F1: . .Q3: 1.23.7 . . . . . . TS: 337-340
F1SR: Q3: 1.24.1 . . X. . . TS: 335 . . . . X
SRF:. Q3: 1.24.9 . .+1 . . .TS: 337-339 . X
F1L:. Q3: 1.23.5 . . .X . . .TS: 342-3 . . . X
-Singapur
F1: . .Q3: 1:42.8 . . . . . TS: 290-294
F1SR: Q3: 1.41.7 . . .-1 . TS: 290-296 X
SRF:. Q3: 1:42.5 . . . X . TS: 296-299 X
-Korea
F1: . Q3: 1.37.2 . . . . . . . TS: 316-320
F1SR: Q3: 1.38.7 . .+1.5 . . TS: 315-322 X
-Suzuka
F1: . .Q3: 1:30.9 . . . . TS: 310-315
F1SR: Q3: 1:33.0 +2 . . TS: 306-313 X
F1L:. Q3: 1:32.0 +1 . . TS: 313 X
-India
F1: . .Q3: 1:24.1 . . . . . TS: 313-19
F1SR: Q3: 1:27.0 +3 . . .TS: 318 X
Test:. Q3: 1:26.0 +2 . . .TS: 317 X
-Abudhabi
F1: . .Q3: 1.39.9 . . . . TS: 315-19
F1SR: Q3: 1.38.9 -1 . . TS: 314-320 X
-Austin
F1: . .Q3: 1.36.3 . . . . .TS: 308-12
F1SR: Q3: 1:34.7 -1.5 . .TS: 311 . .X (wide track)
F1L:. Q3: 1.34.7 -1.5 . .TS: 314 . .X (wide track)
Test: Q3: 1.36.3 . . X . . TS: 309 X
Rfactor >>
https://www.youtube.com/watch?v=MUSJUoej6w0
Telemetry >>
http://oi40.tinypic.com/23m36g9.jpg
Real >>
https://www.youtube.com/watch?v=zWqZAgFFIuk
-Interlagos
F1: (Est.)Q3: 1.11.5 . . . . . . .TS: 309-15 (Est)
F1SR: . . Q3: 1:10.3 . . .-1 . . TS: 312-20 . . X
F1L: . . . Q3: 1.09.8 . . .-1.5 . TS: 327 . . . .+10
Conclussions:
Top speed was 'correct' on 14 of 19 tracks. 3 tracks showed deviation around +-5kph and 2 tracks showed some tendency to exceed +10kph margin (Bahrein and Interlagos).
Rfactor top speeds usually are higher because many sim drivers use KERS only on one of the straights, while real drivers use kers on different places of the track (mainly to help acceleration on low-medium gears). It is possible Bahrein, Interlagos, and other tracks top speed data is distorted due to this fact.
In any case, taking in consideration the majority of correct results for top speed data, which was good for tracks with varied aerodynamic needs as Monaco, Hungary, Austin, China or Monza, our conclussion is that we can consider top speed calibration
CORRECT !!!
Lap times showed higher deviations. One of the causes is there are differences on track layout, width, bumpiness, defined grip, etc,...
The proof is the comparison between lap times for different track versions. For example, two Montreal versions seem to be accurate at first sight, but one of them is 1.5 seconds slower.
Austin is another example. Most used Austin version is clearly too wide, and produces lap times 1.5 seconds faster, while another more accurate version (see video) produces lap times closer to reality.
This is a big disadvantage to make the calibration and also to check its accuracy. However, let's analyse the data.
If we take the tracks ordered by aero setups and showing lap time gap tendency:
Highest downforce setup :
Monte Carlo (x,+1)
Singapur (x,-1)
Hungary (x,+0.5)
Nurburgring (+1)
Medium downforce setup: (-1.5, X , +3)
Lowest downforce setup: Monza (x)
It doesn't seem to exist a rule for deviations depending on aerodynamic needs.
Then, we also classified tracks depending on most common corner types (fast, medium or slow corners) and observed real telemetries to find what type of corners are problematic. However, we couldn't find a rule to apply either.
Tracks with tendency to higher deviation:
Silverstone, India, Suzuka: (+2 seconds)
Interlagos: (-1.5 seconds)
Silverstone and India include some corners of high downforce needs, where real cars are much quicker than F1lrclub cars. For example the first corner of Silverstone wing circuit (286kph compared to 265kph), or some of India fast chicanes.
The simulation was accurate on many of high speed corners, even on the same tracks.. Possibly one explanation is sudden track height changes, which have more influence on rfactor cars than they have on real ones at those particular corners.
However, asphalt grip seems to be the main suspicious. For example, 2014 F1 cars were faster on Interlagos than 2011-2013 cars, thanks to the new asphalt (while 2014 cars were slower on the majority of tracks). Possibly Interlagos rfactor versions had too much grip compared to 2013 asphalt.
Silverstone asphalt is new, too, while rFactor silverstone versions don't feature such difference on track grip for the new sections.
Well, this is all a speculation, and there are a lot of other factors (wind, temperature, asymetric setups, flexible wings, fric systems, etc ).
Average deviation from real lap times is around +-1 seconds, with a correct approach for different tracks as Monaco, Hungary, Spa, Monza. However, it is not the accuracy we were looking for. Probably our next step is define required track asphalt grips, and this will provide a better approach.
Why look for speed accuracy? Obviously we are talking of simulation. We didn't want a F1 car which is 3 seconds faster than reality on every track, or 30kph slower on every straight, or too fast on every corner. Looking for something real is the essence of simulation... and it is good to save time when looking for ideal wings setting !!
Of course top speed and lap times are the result of all car features, but the most important are engine power, tyre grip and aerodynamics.
For the aerodynamics we used drag and lift functions with concavity and convexity. As we can see on J.Katz's "Aerodynamics of race cars", experiments show that lift functions are slightly concave, while drag functions are slightly convex.
We can't obtain F1 2013 aerodynamic data, but probably these curve definitions still work and allow a better approach to different tracks aerodynamic needs, instead of pure lineal functions.
http://i61.tinypic.com/mtmgid.jpg
Next: tyre degradation accuracy report!!
Bye
Rain