Driver pitting with no warning?

my driver had an odd problem this morning.

on the last lap, his pit light came on. i went over to untick it and before i could even choose to unselect it, it turned itself off. I then moved my mouse away and re-selected my kers button (dumping it on the last lap) and then the pit light came on again. i stopped kers'ing and went over to unselect the pit button AGAIN (because I knew he had more than enough fuel to finish) and it turned itself off again. then for a third time the pit light came on and stayed on this time, so I unselected it and he finished the race fine with his fuel still showing "2L" in the little window on 2D.

so what gives? he tries to pit 3 times, and changes his mind twice about it? more to the point why is he trying to pit with that much fuel left......

In our race just finished

http://igpmanager.com/play/?url=race-result/136923#race

Tambay was leading on the last lap, by a goodly margin, the guy playing was on and controlling his KERS, with over 1l left in the tank, and he got auto pitted, without warning, costing him the race win. This isn't the first time it's happened to him either in the last season and a half.

I always keep my pit strategy window open and I didn't notice pitting light there. Can you devs check if it was on or off? I think you should make it impossible to pit in last lap. If that's not possible then maybe a flashing pit light in red colour? It would be much easier to notice.

It doesn't always come on Mikko. I've reported this in my other thread (http://igpmanager.com/play/?url=forum-thread/2114), the driver pits with no warning. Last lap, more than enough fuel, no green pit light and driver just enters the pits.

Atleast this one told me, but in my race now Vettel, once again, maybe for the 15th time tried to pit on the last lap when he had plenty of fuel. In fact to be completely exact, he had an extra litre in the last stint than he did in the stint before, and that was for the same laps. 38L in the third stint, 39L in the last stint, so he was WELL safe.

I spent a fair bit of time today investigating why this may have been happening and came across one bug that is likely the cause.

By design, when the leader is on the last lap, if a driver is low on fuel and will not be able to make one additional lap, then the driver estimates whether or not he will reach the finish line before or after the leader.  If he is likely to reach the line first he would need to complete an extra lap and so must pit.  If he is likely to reach the line after the leader then he will just continue and finish the race behind the leader on the current lap.

The bug I found is in the identification of the leader. The system was originally referring only to a per-lap record in order to identify the leader rather than using a realtime method.  So the above calculation would yeild incorrect results if the leader had switched positions during the final lap of the race.  I have checked against several of the reported races where the leader pitted on the last lap and it does seem that in each case the leading car had changed during the final lap. So I expect this was the cause.

I will apply a fix for this within the hour.  Sorry it has taken so long to address this issue but I expect that is it finally resolved now.

That sounds like some piece of detective work!

Thanks Andrew - I do hope the issue is resolved.
It certainly opens my eyes to how a simple piece of coding (which seems logical until the paramiters change) can have a huge impact.

Good job Andrew!