I recently arrived at the conclusion that the USTA NTRP algorithm is significantly more likely to disqualify players from areas with inherently low tennis participation. In fact, it is intuitively obvious by considering how proper statistical data analysis works.
Match results are the data that drive the calculations of the USTA NTRP algorithm. Large data sets have high statistical significance and lower variance. The vast number of matches that are fed into the USTA NTRP computer each year achieve statistical significance in aggregate. However, the algorithm is also performing calculations on individual players. A player who has played very few matches will have much higher variance in their rating than a player with a lot of matches in their recent history.
Self-rated players are essentially the only people at risk of NTRP dynamic disqualification. Medical appeals, Tournament, and Mixed exclusive rated players conceptually are also self-rated per the USTA League Regulations. Any subtle distinction is in the noise.
Self rated players start out with less than three matches in their playing history. Consequently, their initial performance rating is heavily weighted on their first few matches played. On an individual player basis, this is where the NTRP algorithm will produce the highest variance when calculating the performance level.
Players in areas with a large tennis population are much more likely to play their initial matches against people with a more statistically significant match play history. In essence, the computer doesn’t know anything about the self-rated player, but it knows a lot about their opponent. That makes the calculation less likely to skew to the tails of the curve where a disqualification “strike” is incurred.
Self-rated competitors in areas without a lot of other tennis players are highly likely to play those first few matches against other people who also have a limited match history. In that case, the NTRP algorithm calculates a match performance of two players that it may know next to nothing about. That situation will have much more variance than calculations performed on the general player population at large. It significantly increases the likelihood of a disqualification “strike.”
Players in underserved tennis communities are likely to arrive at Sectionals with a minimum number of matches under their belts. They are also more likely to already have one or more strike against them coming in. Their odds of disqualification are significantly higher than self-rated players with more matches played against a statistically stable population.
As an additional challenge, players going through the self-rate process in underserved tennis communities don’t typically have a rich source of other players to compare themselves to. In the absence of that, honest self-rating errors are more likely to occur.
Unfortunately, there is also another factor that heavily influences the self-rating decision in less populous tennis areas. Typically the lowest NTRP level with enough players to support a league or a team is much lower in rural areas than urban areas. A player might be incentivized to think of themselves at a lower NTRP rating because if they went any higher there would be no opportunity to play at all.
It simply isn’t fair. It is also contrary to the stated objectives of the USTA to make tennis universally accessible to all people. Unfortunately, the inequities don’t stop there. This topic continues tomorrow.
- USTA League Regulations, 2022, downloaded February 5, 2022.
