Skip to content
master
Switch branches/tags
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.

README

The contained files are python scripts designed to process ranking files for sporting matches

The input files are of the form
TEAM1 score TEAM2 score 
with optional trailing values. TEAM1 and TEAM2 must contain no spaces.

The process-* scripts will ignore any line containing a leading #.

rankerfuzz-final.py generates a ranked sort of the teams based on a combination of topological sorting and inference from transitive score ratio calculations.
rankerfuzz-final-1.py does not use topological sorts, and instead iterates to a set number of iterations to achieve a consistent inferred score ratio set, and sorts using that. 

(Both of these files have property flags in ALLCAPS at the top of their source.)

heatmap.py will generate a (HTML table) heatmap for the variation in predicted rank from an ensemble of rankerfuzz* runs (set the number oof runs in the rankerfuzz code to more than 1 and FUZZ to True).

The process-* scripts interface to GNU Octave, to allow Massey style game matrices to be used for linear least squares estimation.
process-for-Octave generates an Avector (which is the matrix of observations) and a yvector (the vector of observed values) for importation into Octave. The yvector contains 3 columns, for scoredifference, log(score ratio) and log(FlatTrackStats style normalised difference ratio).
After least squares fit, the exported beta matrix can be processed by process-from-Octave to give the rankings for all three metrics.

We also now include scripts for importing FlatTrackStats csv-structured snapshots, and octave scripts for generating the ranking matrices.


****
Experimental - contains various test implementations of ranking and rating algorithms which take in FTS data
SRDRank - contains the full SRD Rank (current) algorithm and workflow
Geo_Location - contains the Geolocation tools for FTS data

Graph-analysis - will contain graph connectivity tools for FTS Data, including WebGL Globe
SRDRank2.0 - will contain the 2,0 version of SRD Rank, which will alter the rating algorithm, and also improve the UX for the generated table.

All of this content is licenced CC:BY-NC-SA 4.0 (Creative Commons: Attribution, NonCommercial, ShareAlike 4.0 International - as described in https://creativecommons.org/licenses/by-nc-sa/4.0/ )

About

No description, website, or topics provided.

Resources

Releases

No releases published

Packages

No packages published