This is a great way to experience a full endurance race solo

[ndslpsp]

This is a great way to experience a full endurance race solo. As we all know, the most important thing in endurance racing isn't whether your race pace matches your opponents' in real time. The live standings shown mid-race don't decide the final result – victory depends solely on the total number of laps completed (total distance covered) within the allotted time. You can never be sure of the outcome until the very last moment – and that's one of the biggest charms of endurance racing.

Given that the game engine struggles to synchronise AI tyre wear with the player's, I've shifted the AI's performance onto their final lap totals. For example, I recently ran three 6‑hour races at Imola. I usually completed around 196‑197 laps. So next, I set the AI fuel multiplier to 0.1. They still pit around lap 45‑50 – not because of fuel shortage, but due to tyre wear, as they're forced to change tyres when wear hits about 40%. Fine, let's ignore those details for now. Every car wears differently, their pit stops are staggered, and they're completely out of sync with the player. For a sprint race, that would be chaotic and awful. But for endurance racing, it's exactly what I want. I don't need to stare at a rival glued to my bumper – he might fall behind in the next pit window. I don't need to get frustrated by being overtaken, because I might pass him again soon. Nothing is certain until the finish.

All I need is a field of roughly equal competitors. And yes – with my settings, the AI leader at 95% difficulty completed 200 laps in 6 hours at Imola, the second group 199, the third 198, fourth, fifth, and so on. The first time, I was reckless and just sprinted the whole way – I ended up with only 195 laps, behind the Mustang, a loser even in the fifth group. The second time, I tried a different strategy – saving fuel, saving tyres, reducing pit stops – and managed two more laps, scraping into the fourth group. I think I'll try a third and fourth time – it's great fun. And this is only at 95% difficulty. If I turn it up, the AI will complete even more laps. If I can't race more efficiently and keep improving, I'll get left behind again. I can't imagine what it's like at 100% or even 105%... yeah, it would be hugely challenging and enjoyable – worth attempting again and again.

Under this kind of race, tyre wear, tyre allocation, and everything else can be run exactly like real‑world race regulations. Our tyres wear, their fuel is infinite – and in this strange balance, offline endurance racing comes alive. You might have your doubts – no problem, just jump into a race and you'll understand the fun for yourself.Maybe you'll say, why not just turn off fuel consumption—that's easy, right? But it's not the same. In my setup, the player's side keeps all the normal consumption—that's what makes for a proper race experience.

 

[Sparviero]

But basically it is like playing solo with a target of laps. In my recent experience with hypercars, turn on tyre warm + editing fuel consumption per track (according to player consumption) change the experience, almost solving problems because: 1) AI pits regularly as the player following fuel consumption 2) AI is far more constant in laps time (don't ask me why, but I see AI faster times after 20 laps, not just in first 2-3 laps as usual 3) even if AI did mostly single stint per set of tyres (so it lost about 10 sec every two pits compared to player that do double stints) they don't lost about 15 sec for warming tires, so their race is still alive. I don't know if It works for GT3 too, but I finally enjoy long endurance racers.

Obviously you loss the experience of managing cold tyre but racing is quite safe. However I really don't understand why don't make AI tyres more durable... it seems difficult like go to Mars.

 

[ndslpsp]

But basically it is like playing solo with a target of laps. In my recent experience with hypercars, turn on tyre warm + editing fuel consumption per track (according to player consumption) change the experience, almost solving problems because: 1) AI pits regularly as the player following fuel consumption 2) AI is far more constant in laps time (don't ask me why, but I see AI faster times after 20 laps, not just in first 2-3 laps as usual 3) even if AI did mostly single stint per set of tyres (so it lost about 10 sec every two pits compared to player that do double stints) they don't lost about 15 sec for warming tires, so their race is still alive. I don't know if It works for GT3 too, but I finally enjoy long endurance racers.

Obviously you loss the experience of managing cold tyre but racing is quite safe. However I really don't understand why don't make AI tyres more durable... it seems difficult like go to Mars.

Turning on tire warming has no effect on the AI. Just turn it off. Also, their speed decreases as tire wear increases—they're fastest right after leaving the pits on fresh tires. I always check the details using the 'Results Companion' app, so I'm certain about this.

 

[ndslpsp]

But basically it is like playing solo with a target of laps. In my recent experience with hypercars, turn on tyre warm + editing fuel consumption per track (according to player consumption) change the experience, almost solving problems because: 1) AI pits regularly as the player following fuel consumption 2) AI is far more constant in laps time (don't ask me why, but I see AI faster times after 20 laps, not just in first 2-3 laps as usual 3) even if AI did mostly single stint per set of tyres (so it lost about 10 sec every two pits compared to player that do double stints) they don't lost about 15 sec for warming tires, so their race is still alive. I don't know if It works for GT3 too, but I finally enjoy long endurance racers.

Obviously you loss the experience of managing cold tyre but racing is quite safe. However I really don't understand why don't make AI tyres more durable... it seems difficult like go to Mars.

The AI's excessive tire wear is an inherent engine issue. In rF2, their wear rate is also 3 times that of the player's, and it's very difficult to fix. I'm hoping for an option to turn off tire wear for the AI while keeping it for the player—that would be the best solution. Because with wear turned off, the AI's lap times improve as fuel load decreases, which is more consistent and realistic.

 

[ndslpsp]

But basically it is like playing solo with a target of laps. In my recent experience with hypercars, turn on tyre warm + editing fuel consumption per track (according to player consumption) change the experience, almost solving problems because: 1) AI pits regularly as the player following fuel consumption 2) AI is far more constant in laps time (don't ask me why, but I see AI faster times after 20 laps, not just in first 2-3 laps as usual 3) even if AI did mostly single stint per set of tyres (so it lost about 10 sec every two pits compared to player that do double stints) they don't lost about 15 sec for warming tires, so their race is still alive. I don't know if It works for GT3 too, but I finally enjoy long endurance racers.

Obviously you loss the experience of managing cold tyre but racing is quite safe. However I really don't understand why don't make AI tyres more durable... it seems difficult like go to Mars.

Do you mean that turning on tire warming makes the AI's lap times more stable? If it's for hypercars, I haven't tested them much. I'll give it a try later.Tyre warming definitely takes more than just 15 seconds, and players wouldn’t need to warm up tyres themselves if tyre blankets are enabled, right?Without disabling tire wear, it's impossible for the AI to compete synchronously and fairly with the player. They either pit slower, fall behind you, and then stubbornly block you on track, making it hard to overtake; or they deliberately slow down on the circuit to let you pass, or simply become moving obstacles. If you try to match lap times to find an AI opponent that can keep pace with you in a long-distance race, it's extremely difficult. Play enough endurance races and you'll realize that the only true and accurate measure of fair competition is always the total distance completed at the end of the race.If you find the energy readouts shown by the AI awkward with my setup, you can turn off most of the UI entirely. Keeping only the cockpit dash screen and MFD is more than enough for everything, plus it’s far more realistic and immersive.A racing driver can’t possibly keep checking the virtual energy (VE) readouts of other cars the entire race, after all.

 

[LordF]

@ndslpsp would you mind to share the exact settings you use for an endurance race? It seems to me that you also make changes in the Settings.json file itself.

 

[Sparviero]

Do you mean that turning on tire warming makes the AI's lap times more stable? If it's for hypercars, I haven't tested them much. I'll give it a try later.Tyre warming definitely takes more than just 15 seconds, and players wouldn’t need to warm up tyres themselves if tyre blankets are enabled, right?Without disabling tire wear, it's impossible for the AI to compete synchronously and fairly with the player. They either pit slower, fall behind you, and then stubbornly block you on track, making it hard to overtake; or they deliberately slow down on the circuit to let you pass, or simply become moving obstacles. If you try to match lap times to find an AI opponent that can keep pace with you in a long-distance race, it's extremely difficult. Play enough endurance races and you'll realize that the only true and accurate measure of fair competition is always the total distance completed at the end of the race.If you find the energy readouts shown by the AI awkward with my setup, you can turn off most of the UI entirely. Keeping only the cockpit dash screen and MFD is more than enough for everything, plus it’s far more realistic and immersive.A racing driver can’t possibly keep checking the virtual energy (VE) readouts of other cars the entire race, after all.

Le me better explain what I mean but keep in mind that I report my findings for hypercars so GT3 may be different.

1) Considering stints/no. pits of player, tracks are virtually of two groups: 45+ min stint (most of the tracks), or 60+ min stint (Portimao, Fuji). Le Mans is a class apart, with 12-laps stints. So, in 3H races you did 4 stints/3 pits or 3 stints/2 pits. You have to put the AI in the same windows as you, editing fuel management (I report my guide somewhere). It works greatly, the AI made the same stint as you in longer races (eg. I tried several time 12H of Le Mans and me/AI need to splash at about 30-40 min to the end, so same strategy).

2) This don't solve all the problems because, as you report, AI is not constant in laptime: very fast the first laps, than very slow. If you turn on WEAR but also WARM, laptime are alot more constant. Why? I don't know, probably the AI burn the tyres in the first laps when tyres are still cold (in fact, AI is much faster than player on cold tyres) or probably the physics of AI tyres works differently when warm is on.

2) Moreovere, turn on warming tyres partially solved the most critical problem of AI strategy. The player should not change tyres every stint, but do at least double stint per set, like in real life. This is specially true if tyre warming is off (like in real life) beacause changing tyre mak you loss 12 sec + warming, so about 40 sec in some track to 25-30 at Le Mans. The AI is a bit faster tha you, but the problem is that, usually, don't do many double stint, so in a 3H race it frequently does 1 change of tyre more than you (and at least 3-4 in a 12H race!). So, AI lost a lot of time considering the whole race just because of that, and you easily win even if the race pace of AI is good. The warming tyre eliminate the loss of time by warming tyres by AI, so they just loss the 12 sec for changing tyres, with a not so great impact in the whole race.

 

[ndslpsp]

I've also noticed another fact: AI drivers gradually slow down as their tires wear, but there's no direct correlation between their current remaining tire percentage and their lap times. For example, when their tires are at 20%, they're running 1:30. After adjusting the fuel multiplier, when they pit with 60% tire remaining, they're still doing 1:30. Their lap times only degrade proportionally within each stint between pit stops, and have nothing to do with the specific level of tire wear.

An extreme case: at Le Mans, I set the fuel multiplier to 0.1, and they ran 40 laps before pitting, with tires down to 10% at that point. Yet their lap times were still 3:30 — exactly the same as when running with a normal fuel multiplier, doing 12 laps per stint with tires at 90%.

Based on this situation, if you want the AI to sync with the player, there are only two paths, no other way:

1. Turn off tire wear and adjust the fuel multiplier so that the AI's pace matches the player's perfectly.
2. Turn on wear, lower the AI's fuel multiplier, extend their stint intervals, reduce their pit stop frequency, and aim for a balance in total distance covered compared to the player.

I've taken part in two endurance races. In hours-long races, what everyone enjoys most isn't intense wheel-to-wheel battling — it's clean air, being able to run at your own rhythm without constraints, or tucking in behind an opponent to use less fuel, make fewer pit stops, and cover more distance.

When I encounter a faster opponent, I usually lift off on the straights and let them through — unless there are only minutes left in the race, because defending tooth and nail really just ends up hurting both sides. From this perspective, the pursuit of maximizing total lap count is also a form of enjoyment in its own right, and shouldn't be dismissed just because it's a single-player game.

 

[ndslpsp]

Le me better explain what I mean but keep in mind that I report my findings for hypercars so GT3 may be different.

1) Considering stints/no. pits of player, tracks are virtually of two groups: 45+ min stint (most of the tracks), or 60+ min stint (Portimao, Fuji). Le Mans is a class apart, with 12-laps stints. So, in 3H races you did 4 stints/3 pits or 3 stints/2 pits. You have to put the AI in the same windows as you, editing fuel management (I report my guide somewhere). It works greatly, the AI made the same stint as you in longer races (eg. I tried several time 12H of Le Mans and me/AI need to splash at about 30-40 min to the end, so same strategy).

2) This don't solve all the problems because, as you report, AI is not constant in laptime: very fast the first laps, than very slow. If you turn on WEAR but also WARM, laptime are alot more constant. Why? I don't know, probably the AI burn the tyres in the first laps when tyres are still cold (in fact, AI is much faster than player on cold tyres) or probably the physics of AI tyres works differently when warm is on.

2) Moreovere, turn on warming tyres partially solved the most critical problem of AI strategy. The player should not change tyres every stint, but do at least double stint per set, like in real life. This is specially true if tyre warming is off (like in real life) beacause changing tyre mak you loss 12 sec + warming, so about 40 sec in some track to 25-30 at Le Mans. The AI is a bit faster tha you, but the problem is that, usually, don't do many double stint, so in a 3H race it frequently does 1 change of tyre more than you (and at least 3-4 in a 12H race!). So, AI lost a lot of time considering the whole race just because of that, and you easily win even if the race pace of AI is good. The warming tyre eliminate the loss of time by warming tyres by AI, so they just loss the 12 sec for changing tyres, with a not so great impact in the whole race.

I get exactly what you mean. I’ll run a six-hour race test with the AI while enabling both tyre warmers and tyre wear at the same time — I’ve never tried this combination of settings before.

If the AI can maintain consistent lap times in the latter half of the race, that will already be a fantastic improvement. Even so, I still believe we need to balance out the time gap caused by extra tyre changes during pit stops. The total pit time for every driver is fully logged in the LMU Results Companion tool; it displays the full cumulative pit stop duration for each competitor.

We need to adjust the fuel consumption multiplier to cut down the number of pit stops the AI makes, so their total time spent in the pits ends up roughly equal to mine. Just one fewer pit stop can nearly erase the total deficit accumulated from three extra tyre changes, which will level out the race competition.

That said, this brings us back to the core topic we’ve been discussing: consistency in total race distance. Whether the AI and I make different numbers of pit stops, the race will ultimately boil down to matching total covered distance — as long as our average lap pace and overall pit time are aligned, even if our in-race strategies don’t perfectly sync up. The only caveat is we shouldn’t set the fuel multiplier to an overly extreme value.

 

[Aysedasi]

Over the last two days I've done a 4 hour race at Le Mans. I wanted to replicate as best I could the 4 hour Road to Le Mans race from three weeks ago. I only selected P3s and GT3s and ran a GT3. I turned off tyre wear and damage but left everything else the same. I had an absolute ding-dong of a race for the full 4 hours. I ran closely with two other GT3s while the rest dropped back (but not by huge margins). I won by 0.8 of a second. The guy who I just beat had two small spins and I reckon that apart from those he would probably have won by about 10-12 seconds. I was absolutely flat out for the entire race and so was he. It was quite exhilarating but at the same time it just seemed unrealistic to be racing at the max for 4 hours. I was really pleased with how I drove though, no spins or offs other than a few bad lines into Mulsanne Corner and Arnage. I was taking a second every lap out of the guy from Tertre Rouge to Indianapolis but he then took it all back just through the Porsche Curves to the end of the lap. I want to do more of this stuff (having not driven online for a month now) but making the experience as realistic as possible is obviously the aim. I was so disappointed to abandon my previous 24 hour race with three hours to go but with my opponents pitting every lap (I had tyre wear on) it was just pointless. The only other obvious issue from that 4 hour race was the paucity of opponents - just 14 GT3s and 11 LMP3s - when there were 60 odd in the actual race at Le Mans

 

[aaron0288]

Over the last two days I've done a 4 hour race at Le Mans. I wanted to replicate as best I could the 4 hour Road to Le Mans race from three weeks ago. I only selected P3s and GT3s and ran a GT3. I turned off tyre wear and damage but left everything else the same. I had an absolute ding-dong of a race for the full 4 hours. I ran closely with two other GT3s while the rest dropped back (but not by huge margins). I won by 0.8 of a second. The guy who I just beat had two small spins and I reckon that apart from those he would probably have won by about 10-12 seconds. I was absolutely flat out for the entire race and so was he. It was quite exhilarating but at the same time it just seemed unrealistic to be racing at the max for 4 hours. I was really pleased with how I drove though, no spins or offs other than a few bad lines into Mulsanne Corner and Arnage. I was taking a second every lap out of the guy from Tertre Rouge to Indianapolis but he then took it all back just through the Porsche Curves to the end of the lap. I want to do more of this stuff (having not driven online for a month now) but making the experience as realistic as possible is obviously the aim. I was so disappointed to abandon my previous 24 hour race with three hours to go but with my opponents pitting every lap (I had tyre wear on) it was just pointless. The only other obvious issue from that 4 hour race was the paucity of opponents - just 14 GT3s and 11 LMP3s - when there were 60 odd in the actual race at Le Mans

I suppose that last point goes back to what we were saying in the other thread about not being able to use customs grids… Still hoping they add Michelin Le Mans cup in 1.4, at least that way we can get big LMP3 grids offline.

Try dropping the AI level a percentage or two below where you’re going flat out. That’s what I try to do now so I can incorporate some fuel saving and just making it more realistic in general. As you said, drivers aren’t pushing flat out every single lap in real life.

 

[Aysedasi]

Trouble is, I have a complete mental block (or is that 'mental dislike'?) of fuel saving. I try it but I'm hopeless at it - and more importantly, I don't enjoy it. I'd be a lousy endurance racer in real life.....

 

[ndslpsp]

Le me better explain what I mean but keep in mind that I report my findings for hypercars so GT3 may be different.

1) Considering stints/no. pits of player, tracks are virtually of two groups: 45+ min stint (most of the tracks), or 60+ min stint (Portimao, Fuji). Le Mans is a class apart, with 12-laps stints. So, in 3H races you did 4 stints/3 pits or 3 stints/2 pits. You have to put the AI in the same windows as you, editing fuel management (I report my guide somewhere). It works greatly, the AI made the same stint as you in longer races (eg. I tried several time 12H of Le Mans and me/AI need to splash at about 30-40 min to the end, so same strategy).

2) This don't solve all the problems because, as you report, AI is not constant in laptime: very fast the first laps, than very slow. If you turn on WEAR but also WARM, laptime are alot more constant. Why? I don't know, probably the AI burn the tyres in the first laps when tyres are still cold (in fact, AI is much faster than player on cold tyres) or probably the physics of AI tyres works differently when warm is on.

2) Moreovere, turn on warming tyres partially solved the most critical problem of AI strategy. The player should not change tyres every stint, but do at least double stint per set, like in real life. This is specially true if tyre warming is off (like in real life) beacause changing tyre mak you loss 12 sec + warming, so about 40 sec in some track to 25-30 at Le Mans. The AI is a bit faster tha you, but the problem is that, usually, don't do many double stint, so in a 3H race it frequently does 1 change of tyre more than you (and at least 3-4 in a 12H race!). So, AI lost a lot of time considering the whole race just because of that, and you easily win even if the race pace of AI is good. The warming tyre eliminate the loss of time by warming tyres by AI, so they just loss the 12 sec for changing tyres, with a not so great impact in the whole race.

After careful testing and comparative data analysis, I have confirmed the following (this applies to both rf2 and LMU):

1. With tire wear enabled, AI lap times increase progressively within the same stint, by a range of about 1.5–2 seconds. Additionally, the AI spends an extra 12 seconds on tire changes during each pit stop.
2. With tire wear disabled, the AI's overall lap times are about 1 second faster. On top of that, within the same stint, AI lap times decrease progressively by a range of about 1.5 seconds, and there is no longer any tire-change time during pit stops.Under this setting, both lap-time progression and pit-stop duration are synchronized between the player and the AI, but the zero tire wear makes the race very boring.
3. Whether tire warmers are enabled or disabled has no effect on the above results, nor does it affect the total number of laps completed by the AI.

Given that two games from different eras exhibit the same characteristics in unison, this suggests it is an underlying engine-level trait that is unlikely to be changed.
Based on the above facts, adjusting the fuel multiplier can ensure that the AI completes the same number of laps as the player within the same stint, but their individual lap times can never be synchronized with the player's. They will inevitably either keep pulling away from the player or keep falling behind.