Tyre wear and fuel - Simulating different eras of F1 - Need your help

Hello! (My english isnt so good)
I just want to share some of my experiences about that. There were 3 typical era in F1:
Pirelli and no refuelling - 2011-2013 (in 2010 season the tyre wear werent so significant
Before Pirelli and before no refuelling 1994-2009
and the 2005 season (with one tyre set for all race)

Important! I only examined the speed and strategy for CC cars. I dont talk about human players.

2005
In 2005 the speed of the cars only depended from the actual fuel weight. It can be easily simulated in Gp4. In magic data set "AI Race Performance" (the tyre wear) to zero. From that point the only factor about CC cars speed will be the actual fuel weight. at this season the 3-stop strategies were common, especially in tracks where the overtaking is possible.
I know that "no tyre wear" isnt realistic (Fisichella - Sepang 2005), but this is the way how the pit strategies can be real.

Pre-Pirelli
This can be simulated too (the Gp4 simulates the 2001 season, so it is easy). In that period the driver who can run a longer stint (pit 2-3 laps later, than his rivals) has advantage. My tip: set "AI Race Performance" about 800-1000. I think you should left it on original value. Other people in this community has already tested this function: [docs.google.com]

Pirelli and no refuelling
This was I cant simulate for a long time until I make a test with decrease the Fuel parameter in TSM. And YES! If you decrease it to 200 (or under 1000), the fuel has no effect for the speed of the CC. Important that if you set it to zero, the Gp4 crashes. Otherwise if you decrase the Fuel you also have to decrease the "Fuel consumption AI". If you dont do that the cars will use 4-5 fuel unit in one lap.
I tested it with the 2013 GP4 Central Mod V3 in Sepang (f1Virtual). My settings now:
Fuel 200
Fuel Consumption 350
AI Race Perfomance 6000
So the only thing important is the tyre wear (the AI Race Perfomance). With these setting after 15-16 lap the drivers lost around 1,5 seconds. After the pit stop they are as fast as they can, because the fuel has no effect! I think it is very close to the 2013 season. After that you can play with pit stops (1-2 or 2-3 pit stops).
The only thing that We cant simulate that drivers gets faster as the race comes to the end. The other thing that isnt perfect the mode of trye wear. Pirelli usually good for some laps and after rapidly gets worse. The Gp4 tyre progressively gets slower.



Edited 2 time(s). Last edit at 10/31/2014 12:14PM by somogyi23.

One more think:
the decrease of fuel changed the AI behavior at starts, they are more agressive. I always set AI crash to 1 but if you set it 32 or 64 (common values), it may can cause problems... or maybe not. It needs testing

very interesting to make the 2013 season,keep it up mate good job

Greetings
Alessio


CoM3 oN McL4R3n.....

Settings now for Sepang:
Fuel 200
Fuel Consumption 350
AI Race Perfomance 6000
AI Tyres 6000

I will test this method for other tracks

Great work for the community, keep it up!

__________________________________________________________

"Anything can happen in Grand Prix racing, and it usually does."
"The lead car is unique, except for the one behind it which is identical."

Yeah great work. Would be nice to get this added to the FAQ

---

Races: 163 - Wins: 23 - Pole Positions: 24 - Fastest Laps: 22
Season 9: Constructors' Champions

Trying to simulate different eras is a difficult yet rewarding activity. One has to be creative to find the best approximation; things like the two-compound rule or refueling ban are impossible to simulate. For the modern era scaling down the fuel and fuel consumption parameters is a good first step. Probably the tyre wear values (cc shuffler typically around 3000-4000) in GP4 are close to the post-summer Pirelli tyre wear. In real life, tyre wear depends on the number of high-speed corners, so tyre wear is higher for tracks like Barcelona and lower for tracks like Monte Carlo and Monza. That's why Pirelli selects different tyre compounds for each race, which can be simulated by choosing different tyre types in the magic data. Tyre type 55 is the "supersoft" tyre, while 52 is the "hard" tyre as far as I know. The only problem with the GP4 tyre model is that it does not take "nursing the tyres" into account, but it can be simulated by giving the CC's harder tyres (let them choose the hard tyre instead of the soft tyre).

In the pre-Pirelli days tyre wear was pretty low, but not absent. It seems that there is a nonlinear relationship between cc shuffler and actual tyre wear, with the actual tyre wear decreasing rapidly when cc shuffler is decreased. Values between 1000-1500 seem too low, so maybe 2000 is more realistic for most of the era. However, during the Bridgestone-Michelin tyre war, tyres became increasingly less durable, so perhaps the original cc shuffler values simulate the 2004 season reasonably well.

This needs more testing, but after some races i run, i think that the most important factor isnt the CC-shuffler. For Pre-Pirelli era I would set it around 2000-3000, but if you left the AI Race Performance at original values, the tyre wear still be higher than in 2004 for example.
For Pirelli era, i set CC-shuffler to 6-7000 because about earlier tests, the loss between the first and last lap was the highest with these settings. If you set it lower even if you set it higher, the driver treat his tyre better. For example I run a race in Shanghai 2011 (F1virtual). CC-Shuffler 7000, Ai Race performance 5000. After 15 lap the driver lost 1,5-2,5 seconds, so if another driver come to the pit, he will able to get enough seconds with a pit stop 1-2 lap earlier.
I didnt test the tyre compounds yet, but I will.

Please clarify this: isn't cc shuffler in the magic data the same thing as AI Race Performance in TSM?

CC-Shuffler is the AI Tyres in TSM

Alright... AI Tyres = cc shuffler, but then what exactly is AI Race Performance in the magic data?

Dunno, but it's always 256 so I don't think it does anything important?

=====================================================


Intel NUC 8i3, 8GB RAM, MS Sidewinder Wheel

In the Magic Data Tutorial there are four tyre parameters:

Tyre Wear - 14012
AI Race Performance - 2718
AI Race Grip - 256
AI Tyres - 9373

AI Race Grip is always 256, AI Race Performance is probably the variable I know as cc shuffler, or overall tyre wear. I guess Tyre Wear is the "player" tyre wear, or some tyre wear adjustment, but then I don't know what the AI Tyres parameter does.

Tyre Wear - 14012 - this is player cars tyre wear, I think
AI Race Performance - 2718 - this is the real tyre wear value
AI Tyres - 9373 - this is the CC shuffler. This has an effect for tyre wear, if you set it to zero, the cars will be faster with every lap (with the decrease of fuel and fuel consumption it should be has a result, that the cars do same laps for the whole race). I think its a parameter to set what you want to be important: the actual fuel weight, or the tyre wear. If you set to 1000-2000, it will be realistic for Pre-pirelli era.
But CC shuffler lose its meaning if we off the fuel weight as we did it. So from that point AI tyres and AI performance do almost the same thing. It just need tests about what parameters are ok.
Now I think usually set the AI Race Performance to 4-5000 and the AI Tyres 6-7000. But... if you set AI Tyres to 2000 and AI Race Performance to 10000 may you get the same result. Or maybe not.
Needs more testing.

Sounds interesting, but as I'm using CMagic, it's still difficult for me to translate TSM language into CMagic language.

In CMagic we have these parameters (number and description):
50 "tyre wear" or "Player tyre wear", usually some high (>10,000) value, must be Tyre Wear in TSM
72 "cc shuffler" or "AI shuffler", somewhere between 2500 and 5000, so probably AI Race Performance in TSM
81 "another performance factor" or "AI Grip" , always 256; AI Race Grip in TSM

What parameter in CMagic corresponds with AI Tyres in TSM? Is it 76 "helps decide if a tyre set needs changing"?

hmm.. I own I was wrong. For some reason didnt test if my statement is true.
So after these:
AI Race Performance - is CC Shuffler.

For me this parameter is very clear. I just set it to 15000 in Interlagos. After 9 laps the drivers lost 2-3 seconds per lap (Interlagos is a short track).

I will rewrite the first post about that.

And the AI Tyres may mean: "helps decide if a tyre set needs changing"?

Am I the only one who thinks that GP4 is more suited for the "modern" F1? The lack of a dirty-air effect mimics DRS and the pitstop strategies are mainly driven by tyre wear.

Yesterday, I watched a post-2009-style race in Indianapolis. Tyre wear was the same as in the original game, but fuel and fuel consumption were vastly reduced (to 1/40 of the original value) in order to mimic the refueling ban. As the game lacks a two-compound rule, it is unlikely to see the kind of strategies we saw in 2010, with an early switch to the hard-compound tyre. Running the same tyre type all race long means one has to distribute the pitstops more evenly.

The race started with (Michael) Schumacher on pole and Barrichello next to him. Scumacher immediately began pulling away from his teammate, who was threatened by the McLaren drivers and Montoya (Ralf Schumacher retired after a few laps). While Barrichello slowly dropped back to fifth position, Häkkinen moved up to second position. Ferrari decided to pit Barrichello early in order to regain the lost positions. Häkkinen en Montoya stayed out longer (Coulthard too retired). Schumacher, however, was on the same strategy as his teammate, and he surrendered the lead in lap 24. Häkkinen now was leading the race, but he was delayed considerably by traffic, especially as his tyres began to drop off, which reduced his performance edge over the midfield contenders. Williams reacted quickly and decided to pit Montoya a few laps earlier to avoid the traffic. The trick worked: Montoya got passed Häkkinen after the pitstop. However, Barrichello was back in second place again, but not for long, as Montoya and Häkkinen outdragged him on the long straight.
So Ferrari's ploy to lift Barrichello had failed, as the Brazilian had to pit yet another time. Then, suddenly, Häkkinen went wide in turn 1 and managed to collide with Barrichello. Häkkinen remained unscratched, while Barrichello was out on the spot.
Schumacher was still leading the race, but as Montoya was reducing the gap, he had to pit another time. He now was lagging Montoya by 15 seconds with 25 laps to go. Häkkinen was only four seconds behind Montoya and he was closing in fast. Montoya's tyres were shot and he wasn't even capable of lapping the Jarno Trulli's Jordan. However, his superior topspeed enabled him to ward off every attack. In the meantime, as the pair was losing time fighting each other, Schumacher closed in rapidly. With a few laps to go, he used the slipstream of both Bernoldi (who was two laps behind) and Häkkinen to take over second place. In the next lap, he outbraked Montoya in turn 1 and went on to win the race.

Even though tyre wear was less severe than I had expected, the strategies are closely matched. In theory, the 1-stop strategy is probably slightly faster, but that advantage is quickly eroded away when in traffic. In the end, I think this was a nice race and a reasonable simulation of how the 2001 season would have been without refueling and without rock-solid tyres.

Funny, I just came across an old GP2 guide which listed a 'magic data' element called tyre consumption in the gp2.exe - if only it'd be this simple in GP4...

Here's the link by the way:

[www.grandprix2.de]

After a lot of test with AI drivers I think:
- The effect of the fuel weight can be easily removed. If you set the Fuel to 200 and AI Fuel Consumption to 300 or 350. Even in Monza the difference was sensible, the driver who planned his first stop 10-12 lap later - he can easily overtake other drivers.
- The tyres in Gp4 has some similarity to modern Pirellis. The normal tyre wear is around 4000-4500 I think. With these settings in first 7-8 laps, the cars getting slower only by 0.5-0.7 sec per lap. But after 16-17 laps, they can lose 2.5-3.5 sec (this is driver and track specific).
- The only problem about simulating the 2013 season is that drivers who planning more pit stops they lose a lot of time behind the slower drivers (for example in Melbourne). At a lot of track in Gp4 AI only have 1 or 2 section where they can overtake each other. For example I found that in Istanbul the overtaking is almost impossible.
So if you take a track, and make these settings, and set 2 and 3 pitstops for AI drivers in TSM, you will get almost realistic races.

NOTE!
I always play GP4 with AI Crash set to 1. I hate this random crash system which is absolutely unrealistic.

really interesting mate,so we just have to change those two settings nothing else?thanks for your testings

Greetings
Alessio


CoM3 oN McL4R3n.....