0

Right here are some numbers after week 1 of testing! One other diagnosis will come after week 2. I might be chuffed to use your recommendations!

Notes on numbers

Handiest lap instances clocked on the C3 compound are old to manufacture the three important graphs (lap instances distribution, telemetry comparison and time per half). This due to for the duration of assessments, the F1 live timing app made no distinction between all compounds (C1 to C5). Both C1 and C2 are marked as ‘HARD’ whereas C4/C5 are ‘SOFT’. Honest the C3 compound, the yellow tyre, is abnormal and marked as MEDIUM, permitting consistent comparison between all groups.

There could be now not such a thing as a assumption with regards to gasoline hundreds and tyre degradation. It is a long way exclusively correct-seeking to articulate that plotting a whole bunch of instances per group mitigates the affect of those unknowns. Vitality-unit modes are additionally unknown. This is succesful of per chance per chance heavily affect the comparison in vitality tiny and excessive-flee sections.

Lap instances distributions are smooth with a box keep of dwelling and a violin keep of dwelling. The distribution estimation is made by becoming a KDE, which bandwidth, is estimated with the ‘Scott’ design. In immediate, the violin tries to compare a couple of Gaussian curves on the on hand points. Groups are sorted by easiest time (observed easiest performance).

A telemetry comparison is equipped. Right here, the most arresting laps on C3 of the three easiest groups, plus Ferrari, are proven. The time delta to easiest is plotted along side the tempo hint. You are going to additionally search the division of lap distance in low-flee, medium-flee, excessive-flee (corners) and vitality-tiny (straights). This mapping is the deplorable to calculate instances for the third graph.

In the end the half instances. For every group, the instances spent on each half are averaged and calculated comparatively to the quickest time. For instance, have in thoughts the most arresting lap time of 1: 15.732. If Mercedes takes 22 seconds on common in the low-flee sections and Haas 22.5s, being Mercedes quickest in low-flee the relative loss might be 0.7% ((22.5 – 22) / 75.732 = 0.0066). For this keep of dwelling, I attempted to decrease the thought to be lap instances to the quickest recorded for each group. So handiest +1.5% of the time from the quickest of every group (about 1 2d).

Notes on code

The framework old for this diagnosis is a python kit called Snappy-F1. It is a long way begin-offer and documented (a runt bit). You are going to fetch a contrivance to search out it on Github here. The library fetches and parses files from the known Ergast relaxation api and the f1 live timing files. The knowledge from the assessments is a runt buggy, nonetheless I modified into succesful of derive round a lot of the points. The derive model is the one old for this diagnosis.


Like it? Share with your friends!

0

0 Comments

Your email address will not be published. Required fields are marked *