How often to players playing up pull the upset and win? Interesting Tennis League Stats

In USTA League in the main Adult leagues, players may play at their rated level, but may also "play up" one level.  For example, a player who was a 3.5C at the end of 2018 can play on a 3.5 team but also a 4.0 team*.

An oft discussed question is whether playing up is a good idea or not.  Proponents of it say it give players more opportunity to play and gives them a chance to challenge themselves playing higher rated players.  Opponents of it say it causes the the higher level to become watered down when too many players play up, and leads to pointless uncompetitive matches.

I thought it might be interesting to look at some stats to see what the competitiveness of these matches is and how often the player playing up actually pulls the upset and wins.

For the purposes of this analysis, I'm looking at matches between players that are currently 3.5C and 4.0C rated players playing singles matches in the 18 & Over or 40 & Over leagues and looking at matches playing during the 2019 calendar year.

First, I show that there have been just under 10K of such matches played, 9,941 to be precise as of when I'm writing this.  So that is a pretty good sample size.  Of those, 1,899, or 19.1% were won by the lower rated player.

On the surface, that sounds pretty high.  Nearly 20% of matches between such players are won by the player playing up!  But when you think about it, is it really that surprising?  Even if every player was strictly within the range for their level, you would still have a 3.5 who is a "3.50" play against a 4.0 that is a "3.51" periodically, and you'd really expect the 3.5 to win about 50% of these.  Add to that that players playing up are often played on court 2 singles and that is often where the weaker at-level players are found and the nearly 20% is less surprising.  The consider that some 4.0s are declining and about to be bumped down to 3.5 while some 3.5s are improving and about to be bumped up, and that the "3.5" is probably favored in these, and the 20% and less surprising still.  And then consider there may be a few matches where the 4.0 was intentionally losing trying to manage their rating down ... 😞

Second, it is probably worthwhile looking at how many of these matches actually have the 3.5 rated high now (using my ratings), and we see that 1,255 of them have the 3.5 now rated higher, and so arguably they were supposed to win.  But only 937 were actually won by the 3.5 where they are now not higher rated, so this is probably the real value that should be considered the number of upsets, this being just 9.4% of the matches.

Third, let's look at some details from these true upsets.  Of the 937, 447 of them were straight-set wins for the 3.5, and of these they gave up:

• six or fewer games 220 times
• four or fewer 103 times
• two or fewer 29 times
• one or fewer 8 times
• twice the match was won giving up no games

So when the 3.5 wins, the match is usually very close to competitive, very rarely is it both an upset and a lopsided win for the 3.5.  And while there were two times it was a double bagel for the 3.5, considering all the variables in a tennis match, e.g. good/bad days, injured but complete match anyway, etc., that really isn't that many.

Fourth, what about how competitive the losses are?  Where the 3.5 lost, 1,079 were in a 3rd set, 10.9% very competitive.  But the 3.5 won:

• six or fewer games 5,738 times (57.7%)
• four or fewer 4,157 (41.8%)
• two or fewer 2,080 (20.9%)
• one or fewer 1,052 (10.6%)
• no games 349 (3.5%)

Depending on your definition of a lopsided, 20-40% of the total matches where a 3.5 plays up against a 4.0, it is a lopsided win for the 4.0.  And the 3.5 gets one game or fewer in more than 10% of the matches.

In summary, we see that while about 20% of these playing up matches are won by the player playing up, less than 10% should really be considered upsets, and of those, no more than a few hundred are approaching being a lopsided win for the lower rated player.

What do you think?  Does allowing playing up do more good than bad?  Is it a necessary way for players to challenge themselves?  Or is it abused and far too many players play up leading to uncompetitive matches?

* - Some areas may have restrictions on playing up or how many teams a player may play on in the same league