Opened 13 years ago

Closed 13 years ago

#75 closed task (fixed)

Lc calculation still odd in 1.1.4

Reported by: anonymous Owned by: Stein Tore Erdal
Priority: major Milestone:
Component: FsComp Version: 1.1.4
Keywords: Cc:


Just checked the improvements made in v1.1.4 and have the following observations:

Basic assumption = less area under the distance-time curve means a pilot leaded more time, as he was likely covering distance first.

1) If a pilot lands short his line should be extrapolated not by an horizontal line, but by a vertical line up to the time of the pilot reaching goal last, just as described on the "missing tracklog" example. This way he gets more area to compensate for pilots flying back him but at the end flew in front of the place he landed. As I understood this was the explanation made on the GAP2002 manual.

2) Do not understand the purpose of An factor in the formula, why not just comparing Iv's of all the pilots and get Lc as a proportion of the smallest one?

3) The diagrams in Lc explanation does not have "flying backwards" correction mentioned in the text. Is this actually done on internal calculations?


Change History (5)

comment:1 Changed 13 years ago by Stein Tore Erdal

Owner: changed from somebody to Stein Tore Erdal
Status: newassigned
Version: 1.1.4

comment:2 Changed 13 years ago by Stein Tore Erdal

Type: defecttask

True, this is how it is done. The second drawing in the LeadingCoefficient page may not show this clearly (E is time of last pilot in goal).

I agree and this is one difference from the description and what is actually done in FS. I FS we just use:
Lc = Iv /(1800*D)

For each tracklog-point FS check if the pilot is closer to ES than the prev. point, if so it is added to the dist/time graph with given dist., if not it is added with same dist. as prev. point, ie creating vertical line.


comment:3 Changed 13 years ago by anonymous

1) Great Thanks! 2) So the formula in the text is not what is actually done and the An factor is not used? Are Iv's somehow normalized afterwards for Leading points calculation? is somewhere some information on the formula for getting those? 3) Great thanks!


comment:4 Changed 13 years ago by Stein Tore Erdal

Iv is calc. like this: for each point in time (where getting closer to ES) (dist is in km and time in seconds):
Iv += (time_graph[i].Dist - time_graph[i-1].Dist) * time_graph[i].Time;

This means a pilot using 1 minute interval on the tracklog will get slightly greater Iv than a pilot using 1 second interval (given exactly same track).


comment:5 Changed 13 years ago by Stein Tore Erdal

Resolution: fixed
Status: assignedclosed
Note: See TracTickets for help on using tickets.