UTR Can Be Right but It's Wrong for Kids

UTR's limits make it a poor choice of stat for kids.

Bringing this blog back after a long break. Thanks for reading.

The Kids Love UTR

When tennis kids meet, the first question is, “What’s your UTR?”

I recall a year or so ago, my kid and opponent were walking to their match court. My kid’s opponent asks the UTR question. My kid tells the truth. The opponent lies. Luckily, I hear the lie and with a quick search on my phone showed my kid the truth before they arrived at the court. I say lucky because if my kid believed the lie, there is no way it would not have affected the match.

Kids believe in UTR. They study the UTR of their future and past opponents. They look at UTRs of the top 10 kids and compare UTRs of the pros.

Having a “good” UTR for your age gives kids a huge boost in confidence and respect from other players.

UTR looks serious. All those numbers make kids feel legit. Tennis kids work hard right? We all believe they deserve to feel legit and UTR feeds all that emotional need.

Tennis Programs Love UTR

Tennis camps, clinics, and classes love UTR. I wonder how programs evaluated kids before UTR? Did they use rankings? Most kids don’t have a ranking that means much and the kids with good rankings are probably already too advanced for most programs. UTR really saved the day. The programs should pay UTR for this free service.

When we first encountered the UTR question on a camp registration form, our tennis kid was still playing green ball. To make sure we had a UTR to submit with the registration, we signed the kid up for a bunch of UTR matches. Four matches later, the kid had a UTR. It was relatively painless and not more expensive than anything else in tennis. Not surprisingly the program that was providing the camp also provided the UTR matches.

When good programs use UTR they just use it as a starting point and then move kids into appropriate groups. When not-so-good programs use UTR, they use the rating as a lazy way to jam kids onto a court. Parents don’t question the decision because it’s UTR.

UTR Seems Reliable

It’s true that in many kids matches it seems like a .10 or even .01 of a difference in UTR score will mean a pretty uneven match. In other words, a 5.01 will usually beat a 5.00. A 5.10 will beat a 5.00 6-4, 6-4 or better.

Of course in the pros things are less clear. Witness Jakub Mensík’s (UTR 15.64) recent win over Djokovic (UTR 16.26). A difference that big in the juniors will often mean bagels for the player with the lower rating or would it?

UTR scores mean much less in the higher level juniors. A look at the top 25 12u boys shows some strange UTR discrepancies.

Below on the left is a table with the top 25 12u boys in rank order. To the right of it is the same table sorted by UTR. If we ranked by UTR, things would be a lot different. The highlighted players have UTRs that are “better” than their rank.

There are lots of reasons for this discrepancy between UTR and ranking but with 13 out of the 25 players showing a discrepancy, it looks like a UTR and ranking mismatch is the norm.

Would you be surprised that Borchard leads the head-to-head with Gardality 4-1 winning even their most recent match? Not really because Borchard is higher ranked. So why is Gardality’s UTR so much higher? Here’s the reason as near as I can figure out. Gardality has been playing mostly 14u in the last few months. That has him facing and beating players with higher UTRs. Gardality’s 9.60 is perhaps the right number but then shouldn’t Borchard be at least equal based on their head-to-head?

Here’s another twisty UTR situation. Haris Shahbaz the current boys 12u #10 has a UTR of 7.99. Dmitriy Flyam the #15 has a 8.13 UTR. So what’s the reason for Flyam’s higher UTR? From a deep dive in the UTR site, the only reason I can find is that Flyam has beaten Shabbaz in their two head-to-head matches. Flyam has never beaten anyone with an 8 UTR but he won 2 more games than Shabbaz did in their most recent h2h. And now because Shabbaz is a 7.99, the UTR algorithm makes Flyam’s UTR higher. This situation makes the Borchard v Gardality numbers seem even stranger.

These scenarios are a little complicated right? Here is another weird one.

Thomas Gamble is the #20 12u boy with a 7.59 UTR. Sushant Pothula is #25 and also a 7.59 UTR. Gamble has about 1000 more points than Pothula. Strangely they have both beaten kids with 8 UTRs but their UTRs remain under 8. Also of interest Pothula lost to Harris Shahbaz while Gamble beat Shahbaz and those results impacted each boy’s UTR scores in opposite ways.

These discrepancies are not hard to find. Skim through the rankings and you too can find plenty of examples. Because these discrepancies show up so frequently, you have to assume that UTR is not a very good measure of who will have the most success on the court. A system that’s right 50% of the time is not any better than a guess.

Why does UTR stumble?

I’m no a computer scientist but I have enough software and data experience to know when the rules of an algorithm are not quite up to the task. Here are some obvious reasons why UTR doesn’t add up. First, UTR states that it evaluates the last 30 matches. For players like Borchard or Gardality, 30 is less than 50% of matches played in the last 12 months. Borchard played 141 matches in the last 12 months, 80 in the last 6 months. Gardality played 86 and 44. Limiting which results to evaluate within a larger data set will always produce unreliable numbers especially when your are trying to predict a trend as UTR tries to do. Second, UTR counts games not wins. To me this is a pretty weird version of tennis math and misses the real dynamics of a tennis match. There are so many matches where the winner does not win many more games than the loser and sometimes the loser wins more games. Take a match score like, 1-6, 7-5, 7-6. This kind of score would not be that strange in juniors. The winner won 3 less games than the loser. The winner of that match is a good player with the determination to win tough matches but winning fewer games is basically a loss to UTR. Counting games instead of just counting the win eliminates the truth that great players win tight matches. Third, kids who play frequently are going to have more variation in results. More events means more variation. This is a truth of life in general. Will UTR understand those outlier matches that are just bad days or unusually great days. Do those days represent a trend?

How many matches over what period will show improvement or the opposite? Does a tight win mean more or less than a straightforward win? Do consecutive wins mean anything more than individual wins? These are just a few of the complicated but critical questions that UTR can’t get right.

UTR, An Unreliable Snapshot of The Recent Past

When you are looking at a small sample of data from a limited time period, you call it a snapshot. It might tell part of a story but far from the whole story and you would never base important decisions on just a snapshot.

For a kid’s tennis development, a snapshot can be really meaningless. How a kid did in a tennis match 2 weeks ago does not show the limit of that kid’s ability to win today. Factor in everything that could impact how a kid does on the court and it becomes clear that a UTR does not count for much. UTR is one small piece of information about the recent past. If the kid plays a lot of matches, it is not a reliable piece.

When kids use UTR to judge themselves and their opponents, they are really missing the larger story and limiting their tennis education. UTR becomes a cheat code so you don’t have to understand why you are winning or losing. A kid will think, “I lost because the other kid has a higher UTR.” But the real reasons are far more important.

Do We Really Need UTR for Kids?

USTA hates UTR. If you mention UTR to a USTA leader expect a cold stare. Why USTA has offered WTN as a replacement is beyond comprehension. At least ranking represents the highest value in tennis as a sport - winning matches in succession in a tournament. In fact all sports have this same value. A win is good. Two wins in a row is better. Three in a row is great. Four is a row is awesome. Five in a row or more is amazing.

Becoming a great player is a long process and kids have to push themselves to do uncomfortable things - to make mistakes. But UTR puts so much focus on playing safe. At least with ranking, you don’t lose points when you lose a match. With UTR, every loss can drop your score. The best UTR hack I’ve heard was from a couple of high level 12 year olds who suggested that playing a lot of L7s and then just waiting for the kids you beat to get good UTRs is the way to go.

My suggestion is to tell your tennis kids the truth about UTR. It won’t do anything to help you become a better player. It won’t predict your match results unless you let it. And most importantly, UTR will not predict your future.