I know you meant some kind of program like PunkBuster, but seriously, if someone wants to spend £24 just to get nothing in return theres no point in it.
Let them be a fool, just ban them if you see them, and send a replay to the Devs to deal with.
It could be possible to program insim what does the banning automatically, if:
a) Someone makes unrealistic laptime (For example, more than 0.1s faster than WR)
b) Someone drives over certain speed (Depends what car he will be using)
As you might realize it, this isn't 100% reliable way to keep the cheaters away.
I was toying with doing exactly that last time Gekkibi, along with a more complex idea, but by the time I had the opportunity to actually write it, the exploit was closed I will admit that it's a good, and very simple starting approach.
Exploit was closed? Are you refering to test-patch, or has it been blocked from original Y-patch (Some kind of master server fix-jibberish I know nothing about?)? Because I saw cheating couple of days ago on my demo-server.
PS. Best way to prevent cheaters is the way I did it to my server: I set password and gave that to "those who deserves it". After that the quality of racing has gone sky-high.
Then start to code already.
After you finish it, Scawen will publish a fix. But think it another way around using simple logic: If you will not make the program to prevent cheating, Scawen won't publish a fix.
First I started to type "one second faster", but came to conclusion that it would be (practically) impossible to beat the WR on a racing condition (WR is the hotlap-record, am I right?).
WRs are occasionally beatable in multiplayer because you can use more of the track without risk of invalidating your lap, plus bump-drafting is possible in Multiplayer and not in hotlap mode. I'm sure there are other ways to fudge a world record time without actually cheating in the normal online racing sense of the word.
2 seconds faster than the world record would be "impossible" on blackwood/gti for example.
2 seconds faster than the current as historic rev/gt world record would not even be very fast.
There could be MANY bugs in the second one though.
Think about it.. If you hit a barrior HARD, and go flying VERY fast... 'boom' banned. It could ban many people that don't deserve it.
Two or three people have spoken to me, or passed comment about this, and everyones said the same thing. Who said that these were distinct tests or even that this would be a complete list of what to test for? Gekkibi clearly stated that this wasn't a complete solution.
Anyone capable of writing this program would be intelligent enough to take into account acceleration rather than pure speed and lap time, change in the angle of rotation between polling, etc. etc. There are so many variations and variables that you could use, just from a passive server side monitoring point of view. If you wanted to go whole hog and make a passive client side scanner that talked to a "master server" and the client's LFS installation, that collected and analysed data, then the level of which you could take this is almost limitless.
Making it a single distinct test would be failure on epic levels, but surely that's stating the obvious?
Thinking about what you said about acceleration...
Would it be possible to record the speed at various points in the main 'acceleration zones'. For example record the speed near the start of the BL straight, and then a bit (maybe just a few metres) further on.
A bit like those average speed cameras they have on the road, but instead it would calculate the difference between the two speeds. If the gap in mph is unreasonably high for that particular car, then Ban
'No, I didnt fail 1000 times on making the light bulb, I simply found 999 ways how not to make a light bulb' Thats how the saying goes. Or something like that
I will admit that it's a good, and very simple starting approach.
