I've done some testing, looked at graphs in GPLRA, compared my laptimes, read a few posts, etc.
Not just now, but also over the years.
My gut feeling says the laptimes among the different shifter software rank like this:
1:29.6 - GPLShift without clutch
1:30.0 - original (paddles / sequential)
1:30.1 - Fairshift without clutch
1:30.1 - GPLShift with clutch (threshold value: 90%)
1:30.2 - Fairshift with clutch
In race conditions, you can add 0.1s for anything that uses the clutch because it's not unusual to miss a gear now and then.
If you need data from a dragstrip (1320
):2015 Comparison Replays2018 Comparison Replays
It's nearly impossible to measure at times. The differences often are smaller than driver inconsistency.
It appears that people only see the shift delay times, but don't take into consideration that operating the clutch cuts drive for just as long as GPL's default. When you look at the speed graphs in GPLRA, you're surprised how closely matched it actually is.
Personally, I don't like FairShift that much. Maybe I'm too used to GPLShift, but FairShift's clutch algorithms just seem weird to me.
It's as if you put in a gear and then you're required to press the clutch. I don't think that's the right order.
Making the clutch pedal a toggle switch doesn't seem right, either, but I guess that's just what it does.
I agree that an equalising factor should be mandatory, but not necessarily FairShift in all cases. Enabling the clutch in GPLShift accomplishes the same in terms of levelling the playing field and the driver gets to keep the best driving experience GPL has to offer.