The B Side

Padel in the UK has a ratings problem. And as the sport continues to grow, it is becoming a pretty big one. For something that appears as little more than a number next to your name, ratings now have a huge influence on how we experience the sport. They affect which open matches we join, who we choose to play with, which competitions we enter and, whether we like to admit it or not, how good we think we are. Yet there is still no universally trusted answer to a very simple question: how good are you at padel? Different platforms can give different answers, players can start with inaccurate self ratings and a few strange results can suddenly send a level moving in a direction that does not seem to reflect what is actually happening on court. For a sport that prides itself on being social and accessible, that matters more than we perhaps realise.

The biggest consequence is not somebody being annoyed because their rating has gone down. Put a number next to somebody's sporting ability and they are always going to care about it. The bigger problem is what happens when players stop trusting the number altogether. One of padel's greatest strengths is the ability to find a game without already knowing three other people. You can create or join an open match and, in theory, end up on court with three players of a similar standard. But that only works if those four numbers are reasonably accurate. If they are not, experienced players retreat into private WhatsApp groups, new players find it harder to break into those circles and one of the sport's most useful features begins to disappear.

That is why a ratings problem can become a participation problem. Somebody new to an area should be able to open an app and find a competitive game without already having a network of padel contacts. But if a 4 in Bristol can feel completely different to a 4 elsewhere, or two people displaying the same level are actually miles apart on court, taking a chance on an open match becomes less attractive. When you are paying £15 or more for your share of a peak time or city court, you are unlikely to keep gambling on badly mismatched games. The number is not just a status symbol. It is part of the infrastructure that allows strangers to play together.

It was this question that brought us to Jethro Binns for the second episode of The Alborán Sessions. Sitting opposite us, cross legged and looking remarkably relaxed for somebody about to spend an hour discussing algorithms, ratings and the future structure of the sport, Jethro immediately came across as someone very comfortable with his subject. There was very little corporate language. He would joke about blaming his doubles partner one minute and then, almost without changing his tone, explain the mathematics behind expected scorelines the next. His laid back demeanour almost did not match the meticulous knowledge behind what he was saying. A former professional squash player who reached a career high of 86 in the world, he later became involved with SquashLevels before helping take that experience into padel through PadelLevels.

Behind all of the algorithms and technology, Jethro's argument is actually very simple: sport is more enjoyable when players are evenly matched. That might sound obvious, but it is probably the most important point in this entire debate. A rating should not really exist so you can tell your mates that you have finally reached 4.5. It should help you find a good game. We have almost allowed those two things to become confused. Players talk about protecting their rating, getting their rating back or finally reaching a particular number. Somewhere along the way the tool designed to arrange matches has become a target in itself.

That creates another problem. There is a difference between becoming better at padel and becoming better at increasing your rating. If a system rewards wins above everything else, players are naturally going to work out how to give themselves the best chance of winning. That could mean avoiding stronger opponents, choosing particular partners or finding matches that are favourable to their number. Jethro described the possibility of manufactured levels, where players can select the right matches and keep pushing their rating upwards. You can understand why somebody would do it. Nobody enjoys watching their level drop. But once players start playing the rating system rather than simply playing padel, the number becomes less useful for everybody else.

This is where one of Jethro's strongest arguments comes in. We naturally think that if you win, your rating should go up, and if you lose, it should go down. His view is that the scoreline should be judged against what was expected to happen. Imagine two players who are considerably better than their opponents. The system expects them to win 6-2, 6-2 and they do exactly that. They have won, but they have not necessarily shown that they are better than the system already believed. Equally, if the weaker pair make that match 6-4, 6-4, they have lost but performed better than expected. Once it is explained that way, the idea of somebody's rating increasing after a defeat does not sound quite so strange.

Jethro has experienced the other side of this himself. He previously documented what happened to his own rating after losing four matches with four different partners. His displayed rating moved from 4.42 to 4.12, while the underlying calculation he was discussing represented a 27 per cent change in the measure used by the system. His point was not that four defeats should tell a rating system nothing. They clearly should. It was that four doubles results do not necessarily mean one individual has suddenly become dramatically worse at padel. There was an obvious joke to be made about the partners being responsible and, unsurprisingly, we made it. Jethro was more than happy to go along with it. Underneath the joke, though, sits a serious problem.

Padel is particularly difficult to rate because it is doubles. A score tells us what happened to a pair, but we are trying to use it to rate an individual. Anybody who plays regularly knows how different two performances within the same partnership can be. You can have the best match of your life while your partner struggles. You can barely see a ball because the opposition decide to put you in the fridge. You can be targeted for an entire match. Yet when you walk off court, the result attached to both names is identical. Jethro explained that PadelLevels allows for differences in player ability in its algorithm, but there is only so much that can be learned from the final score. Turning one doubles result into four perfectly accurate individual ratings is incredibly difficult.

Then there is the issue of where those ratings start in the first place. Self rating is almost unavoidable when thousands of new people are coming into the sport, but it creates obvious problems. Someone might say they are a 4.5 because that genuinely feels right to them. They then get on court and discover they are closer to a 3. They have not necessarily done anything wrong. They simply had no reference point. The problem is that their incorrect rating now affects three other people. Those results then affect more people and the system has to slowly work out where everyone really belongs. Jethro described this as the cold start problem and explained that PadelLevels can use input from coaches and club managers when initially setting levels. Sometimes a human who has watched somebody play can still tell you something a score alone cannot.

There is, however, a strong case for the systems already built into the biggest booking platforms. They have something a specialist ratings product has to work hard to replicate: huge volumes of real match data generated by players who are already booking courts and recording results. A rating that lives inside the same platform as the match can be simple, familiar and frictionless. More data also gives an algorithm more opportunities to correct an inaccurate starting point over time. That does not remove the problems around self rating, doubles or mismatched games, but it is an important part of the other side of the argument. The question is not whether the existing platforms have useful data. They clearly do. It is whether match volume alone is enough to create a level players will trust across the whole sport.

Jethro admitted that he has been through that trust problem himself. When he first started playing padel, open matches were brilliant. They allowed him to move beyond playing with the same group of squash players and meet the wider padel community. But as he improved, he had too many experiences where the level shown on somebody's profile did not match the player who actually arrived on court. Eventually it became easier to use WhatsApp groups and arrange games with people he already knew. I suspect plenty of regular UK players will recognise that journey. Once you have built up a reliable group, why risk paying for a game with three strangers when you are not confident their ratings mean what they say?

Then we arrive at fragmentation. Jethro used the analogy of a car needing one steering wheel rather than two. At the moment a player can encounter different ratings across different apps, venues and competitions. Which one are we supposed to trust? Padel makes this particularly important because players do not necessarily belong to one club. You might play at one venue during the week, somewhere completely different at the weekend and enter a competition in another city. Your ability has not changed when you walk through a different door. Ideally your rating should not suddenly mean something different either.

The answer does not have to be one company controlling everything. Booking platforms can continue to manage courts, payments, open matches and the wider experience around playing. Underneath that could sit a trusted rating that travels with the player. A result from Bristol could contribute to the same rating as a competition in Manchester. Different booking systems could still compete without forcing players to build completely separate sporting identities. Naturally Jethro wants PadelLevels to become that trusted rating. He is running a business and is a commercial competitor in this space, so his view should be read with that in mind. But remove the company names and the basic idea still makes sense: one player, one level, wherever they choose to play.

Technology could make all of this much more interesting over the next few years. Jethro talked about AI cameras eventually providing far more information than a final score. A system could potentially know who was being targeted, who made the errors, who hit the winners and how many balls each player actually touched. Imagine being put in the fridge for an entire match and the rating system actually knowing it happened. Cameras are already becoming a bigger part of padel clubs, so the idea is not as distant as it sounds. Suddenly the algorithm would not just know that you and your partner lost 6-4, 6-4. It could begin to understand how that result actually happened. For a doubles sport, that extra information could be significant.

There is no perfect answer hiding in one algorithm. Ratings have to deal with inaccurate starting points, uneven partnerships, tactical targeting, different pools of players and the fact that people behave differently once a number starts to matter to them. The existing booking platforms have scale and enormous volumes of match data. Specialist rating systems can focus more narrowly on the question of measuring level. Coaches and club managers can add human judgement at the beginning. Cameras may eventually add information that a scoreline cannot. The most convincing solution may end up combining several of those strengths rather than pretending one of them makes the others irrelevant.

So yes, UK padel has a ratings problem, and I think it is a big one. Not because somebody's decimal point occasionally moves the wrong way, but because trust in ratings affects whether open matches work at all. If the safest way to guarantee a competitive game is to ignore the number and message the same people you always play with, the system is failing at the exact thing it is supposed to achieve. Yet that is also what makes this quite a fun problem to solve. UK padel is still young. Clubs are learning, technology is improving, platforms are changing and the sport is still working out what it wants its infrastructure to look like. The real test is much simpler than any algorithm underneath it. Four strangers should be able to look at four numbers, walk onto a court and have a brilliant game of padel. Right now, we cannot always trust that will happen. If UK padel can solve that, it will have fixed one of the biggest growing pains in the sport.

The Alborán · The B Sides