The mouse pointer should be visible if you move the mouse, so you can operate the replay controls at the bottom of the screen.
If you don't want to see the mouse at all (for movie making purposes, etc) then you can press SHIFT+F once or twice to hide all text and the replay controls, in which case the mouse pointer is invisible.
Scawen, the mouse is still bound by the limits of the screen even when you're making movies and it's invisible. However, I use 2 monitors and all too often in the process of lifting, moving, and repressing the mouse cursor to move the view further (happens a lot because I record at low resolution), I keep clicking accidentally in the other monitor. LFS minimises and the recording is wasted.
I've mentioned it before as a sort of improvement suggestion but I may as well also highlight it as a bug report, because the "clip mouse to window CTRL+C" feature doesn't stop the cursor at the edge of the screen with my dual-monitor setup, at all.
I use two monitors too, and here "mouse clipped to screen" stops the mouse at the sides until I use Alt-Tab or windows key it works correcty if I use Shift-F4. After that I have to switch it off and on again to work.
I'm not sure if this is the reason, but does the AI stick 100% to the precalculated fuel strategy? Because it's possible that a custom setup changes the actual fuel consumption enough to invalidate the original pit strategy.
(If it doesn't stick to the calculation anyway, then ignore the next two paragraphs.)
That said, another thing I noticed is that the fuel usage prediction doesn't take handicaps into account. If you add lots of extra weight for example it's very easy to make it use considerably more fuel than estimated.
Now I don't know how hard or easy it is to change the algorithm for calculating the usage, but a rather easy fail-safe fix would be checking each lap if there is enough fuel left to last two more laps (taking the actual average fuel consumption as base) and if not simply instruct the AI to make a pitstop ASAP.
Which car did you use? And was it using player setup or its own setup?
In fact this is nearly what it does, only it makes sure there is enough fuel to last 1.5 laps (not 2). I guess this may be the problem, on a track where the split point is just less than half way round, there is a chance it will say, "I have enough for 1.5 laps so no need to go in to the pits this time".
If this is the case then it can be fixed by looking at how far round the track that split is. It seems likely that this is the problem but I was hoping for a reproducable case to verify the fix.
I'd perfer to do it more accurately rather than going for 2.0 laps, otherwise I imagine there is a case when an AI could unnecessarily fill up with less than two laps to go, although it could reach the finish line with its current load.
Scawen i just bought the driving force pro wheel can you add to vibrate more when i'm in the grass because in the real world when i get into the grass whit big speed the wheel is going crazy on the road too vibrates in the real little
I don't really understand the reason behind taking 1.5 laps as base? Lets say your fuel lasts for 1.6 laps as you cross the line, so you decide not to pit yet. Then one lap later you check the fuel again, now your fuel lasts for 0.6 laps. Whoops, no way to reach the pits after that.
Wouldn't that basically be a 40% fail rate on the fuel checks (every time it turns out you have 1.51 - 1.9 laps left when doing the check, assuming that you can somehow creep into the pits with 0.9 laps worth of fuel)?
Regarding pit optimization, wouldn't a simple check on "if (laps_to_go < laps_possible_with_current_fuel_amount) then ignore pitstop" do the trick? Or is this a question of saving the CPU cycles?
The reasoning is because it makes that decision when it passes the last split. It has to make the decision before the pit lane entramce and the final split is always before that point. I guess as I was coding it, I was thinking about tracks with two splits, so 1.5 laps would cover all. It just doesn't cover the case when there is only one split and that is less than half way round.
Anyway, If anyone finds a way to reproduce it, do let me know. I tried what you said BOSCHO, an FXO using the AI setup and also RACE_S - in both cases it did pit. This is on a 2 hour race at AS National. I've made a super fast version, I guess I'll send some more cars round there and see if any break down.
But can you pleace make that everybody can download it on LFS ? like Y or so. then i only got the Y version and read i need the Patch Y19.
I mean a Patch update on the server.
Please update that
Would be very nice