just thought of a new charts for LFSWorld

a chart that will show you how does your PB on certain combo stand in comparison with all other racers, so:

x: lap times
y: number of drivers
(both data with some clever aggregation of course)

the graphical outcome would be similiar to the History charts in Racers&Hosts online section with, for example, a red line in the place where you laptime is, the chart itself could be opened by a small button next to your PB

there was quite a similiar suggestion for hotlaps once (yes did the search ) but: hotlaps are not representative, only a small fraction of drivers upload them, they are selected, and so on... and PBs are a huge database, it would give you the info not only how far are you from the WR, but how do you compare with the average laptime achieved online on that combo

the SQL performace issues could also be solved quite easily by preparing the global data let's say weekly instead of performing a full search for every chart call, but that's a complete different story...

the real question is if anyone, except me, wolud like to compare his PB not only with the WR, but also with the whole PBs data, on a single chart

The Personal bests are just orientation and it may contain illegal times / cheating.

well, I could be tempted to make a plain comparison against other racers' PBs. In fact the old lfsworld had that feature.
But charts for PBs to compare your own times to I'll not do, because indeed PB times can be cheated, or influenced by wind / shortcutting / bumping / drafting / etc so charts are useless, but plain comparisons with other racers of your own choice might be nice to have again.

Quote from Victor :

PB times can be cheated, or influenced by wind / shortcutting / bumping / drafting / etc so charts are useless, but plain comparisons with other racers of your own choice might be nice to have again.

hmmm, the question is if the amount of the all the types of "illegal" PBs is so big to influence the overall result? will it really matter in showing if you are within 5%, 20% or 50% of the fastest drivers on a combo? if so, and you should know the best how big the error would be, than it is really not worth it, but if it could still give some kind of image, to see it on the median chart then maybe?

comparisons with other racers also would be great, but IMO, even more affected by the problems you mentioned, when you compare to "illegal" lap one to one then it is 100% of your data, if you compare with the whole database, then the "illegals" are only a certain percent of your data...

Some cheats to get better pb's, but don't forget cruising / drifting servers: I think they compensates the overall score...

Same goes for restricted cars like UFBabyR, 3%-FXO or GT2/GTR-class

well it doesn't really matter if someone did 1000 of laps with a restricted car or on a cruise server.. because we are talking about PBs only here... and if someone really did his PB with a restricted car? than this is his proper PB and it should be counted as one

this is only statistics, and that's how it works

This won't be representative anyway if ALL PBs are included since some people drove just one lap on a combo and have a PB of 60 minutes which will drag the average down.

Quote from zeugnimod :

This won't be representative anyway if ALL PBs are included since some people drove just one lap on a combo and have a PB of 60 minutes which will drag the average down.

this has nothing to do with average values, it's about distribution, range, deviation...

made a small "example" of what I had in mind, because I realized it's quite hard to get the idea without any statistics background

so there is a sample chart in the attachment, the red colour is your PB on a combo

I think it would be most interesting for newer and "average fast" drivers

Sorry for nothing to do with this(I didn't want to create another thread) but i want to suggest something for LFSworld. Let it save the BEST first split you do, saves the best sectors.
And you can check best possible laptime.
I know lots of insims have this, but what if you go on a server, set lots of good splits and then go another server and set only the first split, then you cant check pos laptime.

Codes:
/w pb 1split
/w wb 1split (wb=world's Best)
/w pb poslaptime
/w wb poslaptime
(This would be nice)