After running a tournament, it would be useful to be able to use the results of the tournament to classify new samples.
This is necessary so we can compare the new, unseen data to data with a known ELO.
There is a efficiency/accuracy tradeoff here... a more accurate ELO estimate requires more matches against known data
After running a tournament, it would be useful to be able to use the results of the tournament to classify new samples.
This is necessary so we can compare the new, unseen data to data with a known ELO.
There is a efficiency/accuracy tradeoff here... a more accurate ELO estimate requires more matches against known data