The ultimate aim of SquashLevels is to assess the level at which every player on the system is playing after every match they play and give them a number - their level - to represent that.

The level is mathematically comparative between players such that if you are playing twice as well as your opponent then you will have twice the level and this applies all the way from a beginner's first competitive matches to the top pros. This should apply wherever you are, however good you are, whoever you played and whenever you played!

This allows us to plot graphs, compare players, predict results, set goals and even compare players from different eras! Although it's all comparative, it's fixed to a specific level at a specific time so it can actually be treated as an absolute figure. That's really important so that when a player knows what level they are, they know what that means. These are the sorts of levels we find. Numbers very approximate as clubs, counties etc. vary considerably!

• Beginners - (< 100)
• Leisure players (50 - 300)
• Club boxes players (200 - 2000)
• County league players (500 - 3000)
• Top county league players (3000 - 10,000)
• PSL (10,000 - 30,000)
• Satellite PSA (20,000 - 40,000)
• Top PSA (30,000+)

With the vageries of human behaviour, millions of matches over decades of time, it's a complex task!

Levels are calculated for every player after every match and are based on:

• The player's level before the match.
• Their opponent's level before the match.
• The actual result compared to the expected result. Given we know the player's levels before the match we can predict the result if both players played as expected. If the actual result shows that one of the players played better than expected, their level will go up a bit and their opponent's level will go down a bit.
• Maths - to determine how much their levels should change if there was no damping or behavioural modelling. PAR is easy (11-5 is about twice as good), English scoring less so. We use a combination of points scores and games scores to assess the result. The overall goal is that if you are twice as good as your opponent then your level will be double theirs. This works all the up from beginner (<50) to top pro (>50,000)
• Weighting - the more important the match (e.g. a tournament) the greater the weighting. This allows you to play a box match without having too much impact on your league standings.
• Damping - there is a reasonable amount of damping dialled into each match trying to get the balance right between wild swings and slow progress. The algorithm does its best to reward every player for a good result and, in consequence, an appropriate level reduction if the result isn't so good. The intention is that a player's level is a reasonable assessment of their current playing level within a match or two.
• Behavioural modelling - as it turns out, not everyone puts 100% effort in every match and that’s down to behaviour. There are many other cases too where player behaviour defies the maths and, based on the analysis of 1.6 million results on the system, we’ve built an extensive behavioural model that allows us to predict and make use of these behaviours. See the calibration FAQ for more detail on what behaviours we model.

We can work with game scores only, making assumptions around the average 3-0 result (based on our analysis of real 3-0 match results) but we can only use averages so it takes a lot more results for the levels to become accurate. Not all 3-0 results are the same, obviously.

They're really very different but if you order players by their level you will get a ranking and you can compare that with the latest tournament based rankings and usually see the two lists looking a bit the same but not the same. This can cause confusion and lead to the question; 'Why do we have two different rankings?!' It's a fair question!

Level based rankings - which is what SquashLevels generates - are based on a player's current playing level and (as per the section above) that's entirely based on their actual results against their actual opponents. It doesn't matter who the opponent is (it's just their level that counts) or what tournament it is and not which round they are playing in.

Tournament based rankings are based on points accumulated from progressing through the rounds of tournaments over a number of tournaments. Each round is worth more points (usually double) and the whole event is usually scaled so that the more prestigious events (with the better players) are worth a lot more points. Usually, the rankings are based on a points based algorithm such as the average of your best 10 tournaments over the last 12 months. In this case progression through tournaments is everything - in fact not progressing is often worth nothing!

It's interesting to compare the two methods:

• The points gained at tournaments are exponentially geared to the players that win and progress through the rounds. The winners of the tournaments end up with large numbers of points, whereas those that get knocked out early get very little - or nothing. So when the points are added up over a number of tournaments it's very clear who the winners are as they have masses of points. This is a very good method for finding out who the very best players are - and so will always remain the official ranking method for the top players.
• A lot of squash players don't even play tournaments. Or, if they do, it's the odd tournament run at their club once a year. If you need to play in 10 or more tournaments in a 12 month period and progress through at least one round to score any ranking points then that immediately reserves the tournament based rankings for the best players - just the top few in any pool. It's not inclusive. What you typically find is that the tournament based rankings are spot on for the top few players and then go a bit random after that. Those next few players are the ones able to/prepared to travel around and put the time and effort in. Fair play to them but that doesn't necessarily make them better players than those who don't.
• Also, as SquashLevel's levels are based on current performance and it takes a little while for the best players to emerge at the top of the tournament rankings (because they are averaged over 12 months), the levels based rankings are a bit of a look into the future and more likely what the tournament rankings will look like in about 6 months. Very interesting...

So in summary:

• Tournament based rankings are focused on the top few, averaged over a long period of time.
• Level based rankings are inclusive, current, dynamic and a predictor of future tournament rankings.

This means you have a provisional level which is simply that you have 5 or less results on the system. Your level should be reasonably close after 5 matches but another important point is that while your level is provisional, the system will adjust your level after your matches but not your opponent (unless it's a county league or tournament match). This allows you to get some games in against a few players while your level is settling without impacting their level.

We take the accuracy of SquashLevels very seriously so always want to know about cases like this. There are a few common reasons:

• There could be some duff results such as incorrectly entered scores or injury results messing things up. Have a look at the history of anyone who seems a long way out and check their results. If you see any anomolies please let us know and we can correct them.
• You may have a lot of players with a very similar level so a small change in how they play or an unusually good result can have quite an impact on their position in the rankings. We can't do much about that - that just adds to the spice and the motivation to play every point!
• Some of the players may only have a couple of matches in their history and it takes a while with weighting and averaging for their level to become accurate. Three is usully enough if they've played other players with accurate levels but if everyone has only got a match or two to go on it can be a bit random to start with. This can happen if a new league or club has just started uploading results. If it's bad or you're impatient to fix the ranking order then the starting levels of any player can be overridden. Once there are enough results to go on the system takes over but to get you going just identify a club or county admin and we'll set them up to be able to override the starting levels.
• Maybe some of the matches recorded weren't truly competitive. A results based system can only go on the results so the rankings will give you the order of how well everyone is playing - not necessarily how good they are. The more competitive the matches, the closer those two will get.
• You may have some incorrectly merged players. There's a fairly continous merging of duplicate players going on and occasionally we merge two players who are actually different. That would have the effect of merging their playing history under a single name and put their level and ranking out. Check their player histories and if this looks to be the case then please get in touch and we can sort it out.
• Sometimes we study rankings that don't look right, based on our gut feel, but when we look at the results we realise the rankings are probably right and we're just running on old heresay... If the results are right then the rankings are probably right too.

If you feel that a player is in the wrong order (by, say more than 10% from a relative level point of view) then please get in touch and we'll take a look. The system is no good if it's not accurate and that's our holy grail!

The level changes are damped quite a bit but they do still go up and down by a few percent each match based on your results. If you're going to be using them for tournament seedings it's a good idea to damp them even more so they are more stable. The 4 match moving average level for a match takes the level from each match and averages it with the levels from the 3 previous matches to provide a heavily damped level for that point in time.

For the rankings, this 4 match average applies to the 4 most recent matches.

The level is recalculated for the match just updated but if that's not the most recent match then your latest level will not be changed. Level history is re-calculated every night so just wait until the morning and it will all be sorted.

Sometimes you can see that your level has gone up by, say 5%, but when you look at your actual levels on your history page it hasn't. It may even have gone down!

The most likely cause of this inconsistency is that someone may have updated the results (maybe there were corrections) since the nightly processing was run so now one or two of the results aren't consistent. Level history is re-calculated every night so just wait until the morning and it will all be sorted.

If you click on the level change text in your results history, you will be taken to a page that explains how that level change was calculated. It is the actual level change code that runs in this page, except it's in 'verbose' mode so you can see what's happening (which is why it's a little cryptic). If you still can't see what happened please get in touch and we'll explain it.

This is done automatically by the system and is based on your first few results. As long as someone somewhere in your league has a level, the system can work out everyone else's level based on the chain of results - even if that means working backwards in time. The first pass at this gives an approximate set of levels but this is fine tuned every night so after a week or so of your first results going in, your initial level will be accurate and stable.

Your initial level can also be adjusted by the inter-league calibration process that runs every night as that's how whole leagues are adjusted. These adjustments are normally very small but can be significant over time. Once the leagues are calibrated, they go stable.

We do need two or three representative results within a period of a couple of months to get you going so it can be less accurate if, say, you have one result and then nothing for a few years after that. In these cases it's probably worth a little manual adjustment. If your initial level looks wrong, please get in touch and we can take a look.

As well as having a level, you have a level confidence. This is a figure used to express (as a percentage) how much confidence the system has in your level - 10% for new players, 70% or more for established players. It is adjusted as follows:

• It goes up when you play another match.
• It goes down if you haven't played for a while.
• It goes down if your result was unexpected.
• If you are new, or haven't played for two years then it's set to 10%.

The effect it has is that if your level confidence is lower than that of your opponent then your level will move more quickly than theirs. This allows the system to adjust the levels of the players who's level might be inaccurate more quickly with limited impact on their opponents. It also allows fast moving players to move fast!

SportyHQ is a comprehensive tournament, league and membership management system and has its own ranking system based on 'SportyHQ points'. For those familiar with SportyHQ points and wondering how they compare with SquashLevels here's a summary of the main differences:

• SportyHQ points are based on the 'matrix' system which uses game scores and their opponent's SportyHQ points value to decide if they played better or worse than expected. Points are awarded based on a matrix and are additive. You can get anything up to a 30 point adjustment from a match though it's typically much less than that - 1 to 5 is common, 10 is quite significant.
• SquashLevels uses points scores as well as games scores for greater accuracy and a more mathematical approach to a player's level such that someone playing twice as well will have a level that is twice as high. So levels are more of a ratio than additive. I.e. a 2000 level player will be playing twice as well as a 1000 player whereas the equivalent for SportyHQ is a difference of around 275 SportyHQ points. This also allows results prediction as you'd expect someone with twice the level to win their games by around 15-7... Of course, most humans are less predictable than that but it's good fun!
• As SquashLevels takes points scores into account it is more responsive and dynamic than the SportyHQ points value. This allows for fast moving players such as juniors or those just simply those getting better quickly. This also requires that we take player behaviour into account such as when a good player gives their lesser opponent a runaround.
• A key part of the SquashLevels algorithm is calibration. This compares pools of players such as between clubs, regions and even countries and makes continued small adjustements to ensure that a 1000 player in one club or region is equivalent to a 1000 player in another club or region. This is performed globally and a unique feature of Squashlevels.

We ran a comparison between SportyHQ points and SquashLevels level (based on one SportyHQ region) to create a mapping between the two systems though bear in mind this is approximate as SportyHQ doesn't perform calibration, the absolute values may vary from region to region.

Yes! As we've rolled out SquashLevels in Australia we've had more people look at and check out their rankings and had some feedback that they're not quite as expected! We've investigated this and found two main reasons:

• SportyHQ's rankings are very highly damped. As players ebb and flow in terms of their playing level, SportyHQ remains unphased with their points value only going up and down by a point or so each time. It's certainly stable but restricted when it comes to matching the levels of players that are really changing. SquashLevels is more dynamic and tries to assess a more current playing level. It's also damped but nowhere near as much so you'll find plenty of cases where a player seems to be in the wrong place in the rankings (compared to SportyHQ) but actually they're playing at a different level and, for now, they are in the right place!
• The COVID lockdowns have meant that some players haven't played in over a year and are only just getting back into it. Some players have stayed keen and fit whereas others... haven't. This has really shown up in the first first matches back since February as some players have seen their level rocket whereas others have gone the other waay. With the dynamic approach that SquashLevels has to tracking level, this has resulted in much more player movement than we think is appropriate so we have decided to counter this COVID effect with a lot more damping. This has caused an adjustment to everyone's level in Australia.

We won't be asking clubs or states to reset or re-enter any starting levels. We have developed a mapping between SportyHQ points and SquashLevels level and re-used the rankings as defined by the clubs and states from four years ago for SquashLevels to use as a starting point. SquashLevels has then used its own algorithm from there to get to your current level.

We do have a cunning tool for determining what a player's level should really be irrespective of algorithm, points value or level. We are able to compare any two players based on their results with their opponents, and the results of their opponents against their respective opponents and so on until a linkage is found between the two players. This might sound tenous but there are often hundreds of such linkages between two players in the same league, pennant or tournament and with a bit of good old-fashioned weighted averaging we can get a really good feel for the difference between two players. If you think someone is still in the wrong place in the rankings we can run a player compare and see where they should really be.

Please see the SportyHQ compare FAQ above to learn more about the mapping between SportyHQ points and SquashLevels level.