League 2 Analysis - Plymouth and AFC Wimbledon

With most teams having played 17 games this season, we are over a third of the way through the 2015/16 League 2 season. Plymouth Argyle are currently setting the pace on 38 points and have a five point lead over Oxford United, with Northampton another two points further back. Pre-season title favourites Portsmouth lie in fourth place on 29 points, while at the bottom, Dagenham and Redbridge seem to finally have regressed to the level that their underlying statistics seem to suggest as they lie rock bottom. Just above them are Yeovil, while York, Newport and Stevenage are just outside the drop zone and the two relegated teams are likely to come from those five.

Let us look in slightly more detail at Plymouth. It is difficult to find any underlying statistics that support their current league position. Their total shots ratio of 53.7 is the eighth best in the division, while their shots on target ratio of 54.0 is also eighth best. Their expected goal difference is 3.0, putting them seventh in the league. None of these statistics seem to suggest that they should be five points clear at the top of the table. The biggest area in which they seem to be overachieving compared to their underlying statistics is in defense. Their 12 goals conceded is the best in the division, but this is well below the 22.0 goals that they are expected to have conceded from the shots that they have given up. This overachievement is backed up by a PDO value of 114.7 for the season so far.

We can see how unlikely Plymouth's current overachievement is by using the shots data for the season so far to run 20,000 simulations. The chart below shows the distribution of the expected points from those simulations with their current total of 38 points marked in red.

As we can see, Plymouth gained 38 points in just 61 of the 20,000 simulations, or just 0,3% of the time. Their expected points are much lower than their actual total at 25.5, which is still a very decent total and would put them on the verge of the playoff places. Indeed, in 11.8% of the simulations, they are currently lying in the top three and are in the playoff places in 40.7% of simulations. However, they are only top in 2.4% of simulations.

Based on this, we might expect Plymouth to start dropping down the table in the coming months unless they can improve their underlying statistics.

Looking at the underachievers and we find AFC Wimbledon stand out. Currently sitting in 9th in the league on 26 points, their total shots ratio of 61.8 is second in the division, while their shots on target ratio of 60.7 is fourth. Their expected goal difference of 10.8 puts them third in the division. While Plymouth were hugely overachieving defensively, AFC Wimbledon have the opposite problem. Their expected goals conceded of 13.8 is the second lowest in the division, but their actual goals conceded of 23 is letting them down. Whether this is bad luck or symptomatic of poor keeping or defending is unclear, but one would expect this to regress toward the expected goals in the future.

AFC Wimbledon can consider themselves quite unfortunate not to be in the playoff places at the moment. In the 20,000 simulations, 17,770 of them had AFC Wimbledon in the top seven in the table, while they were actually in the automatic promotion places in almost 60% of the simulations. Indeed, they were top of the table in just under a quarter of all the simulations.

Just as we might expect Plymouth to drop down the table in the coming months, we might also expect AFC Wimbledon to move up into the promotion reckoning.

Transition to Serena 2.0 - A Review of 2015

It was almost an unprecedented year for Serena Williams. She became the oldest woman ever to have won a Grand Slam title, she completed the Serena slam for the second time in her career and she lost just three matches all year, albeit one in the most important match that she played. She continues to dominate the women's game to the extent that her current run of 143 weeks at number one is only behind the great Martina Navratilova and Steffi Graf and you would certainly not rule out the possibility of her completing the additional 44 weeks needed to pass Steffi. Whichever way you look at it, she is the undoubted number one in the ladies game.

This time last year, I wrote an article looking at the stats behind Serena's game and made the claim that her overall game was in decline. Whilst I did emphasise that she was still comfortably the best player in the world, many people jumped on this questioning how Serena could be in decline and still beating every other player. Certainly, on the face of it, Serena's results in 2015 do not seem to back up the idea that she is in decline.

Recently, I read an interesting article about Cristiano Ronaldo and the way in which he was adapting his game to mask the effects of physical decline. Age catches up with every professional athlete in the end, no matter how good and how well they train. However, some of the smarter athletes are able to tweak their game to extend their run at the top of the sport. Combining the stats from 2013, 2014 and 2015, I shall look at whether Serena is adapting her game to overcome the physical effects of ageing.

Basic Statistics

Let us first look at some of the basic statistics. As with last year, I shall focus purely on Serena's strongest surface, hard court matches, for this comparison.

The first thing we notice is that Serena is winning more on her first serve than in the previous two years. Indeed, a nearly 2% increase is a significant rise and shows her domination behind that first serve. This ties in with the increase in the number of aces per game that she is serving. For the second straight year, we can see an increase here, suggesting that her first serve is working better than ever before. Interestingly, her points won on second serve has actually decreased this year and has dropped to below the level from two years ago. Perhaps as a result of this drop on second serve, we can see that she is facing more break points per game than in either of the past two years - this is something we shall look at in more detail later.

On the return side, we can see that she is winning more on return than last year, although still below the 49.7% that she was winning in 2013. However, this has corresponded to creating more break points per game than in either of the previous two seasons. Maybe this is an indication that she is focusing on specific return games - if she falls behind in a game, she is happy to conserve energy for future games.

Looking at the data, we can see that in 2014, in return games where she went 30-0 down, she came back to win 23.0% of those games. In 2015, this has dropped to 18.3%. There was a similar drop in games from 40-0, falling from 14.3% in 2014 to 10.2% in 2015. We can also look at the proportion of return games in which Serena fought back to Deuce from 40-0 down. This has also dropped from 27.1% in 2014 to just 16.9% in 2015.

Based on this, there certainly seems to be some merit to the argument that Serena is happier to let return games go if she falls behind to conserve energy for the rest of the match. However, the increased break points created per game statistic suggests that when she does get into a return game, she goes in for the kill. However, how do these stats convert into what really matters - winning games?

We can see that, despite the increase in her points won behind the first serve, Serena is actually winning fewer service games than in either of the past two years. However, the proportion of return games that she is winning has increased from last year, although it is still well below the incredible 50.3% that she recorded in 2013. Based on this, we might surmise that, while she is winning more points behind her first serve, she is making fewer first serves. Again, the stats back this up with her first serve percentage dropping from 59.9% in 2013 to 58.2% in 2014 to 56.5% in 2015.

It would seem that Serena is going for more and more on her first serve, hitting more aces and winning more points, but at the expense of missing a few more first serves and opening up her second serve to attack from her opponent.

This table helps to back up this idea. She has improved her non-ace first serve points won from last year, which is undoubtedly helped by plenty of big first serves that are resulting either in nonreturnable serves or looped returns that she is able to put away easily. However, the boost that she received in her non-DF second serve points won last year seems to have disappeared, suggesting that her second serve is becoming more susceptible to attack.

So, the broader top level statistics seem to suggest that Serena's first serve is working as well as it ever has. She is continuing to develop what is arguably the greatest serve in the history of the women's game into an even more dangerous weapon. This is important as it provides plenty of short and easy points, allowing her to conserve energy. We can also see a shift to targeting certain return games based on how the first couple of points play out - if she goes behind in a game, she is more willing to let that game slide and conserve energy for future return games.

Let us now look in some more detail at certain aspects of her game. To achieve this, we shall use data from the Match Charting Project on TennisAbstract. One caveat to this section though is that this data is not complete - it contains data from 12/32 matches in 2015 and 11/46 matches in 2014. This obviously means that we may miss a few things or that the data can be overly affected by the quality of opposition, but it is still a valuable source of insight and should give us a few ideas of how her game has changed over the past two years.

Serving

First, let us look at her serving on the key points - break points, game points and points at deuce.

Firstly, we can see in the first column that, while she may be winning slightly fewer game points than last year, the points won on break point and deuce are both up, driven significantly by a big rise in the percentage of aces that she is hitting on these points. In 2014, she served an ace on 11.3% of break points, but this has risen to 17.0% in 2015. Similarly, at the important deuce points, she has gone from serving an ace on 10.5% of points to 24.7%. That is quite astounding - hitting an ace almost one in every four points at deuce gives a huge advantage to Serena in those situations. We can also see a rise in rally winners and rally forced errors at deuce, suggesting that she is really going for it on these points and, in 2015 at least, going for it successfully.

Let us go into even more detail on her serve.

We can see here how she directs her serve, broken down by first and second serve as well. On the deuce court, we can see a big drop in the number of serves aimed down the middle, both on first and second serves. This is visible particularly on her first serve, where she has reduced it from 12.0% of serves down the middle to just 3.7%. Instead, she seems to be aiming these serves both wide and down the T - lower percentage serves, but with a greater chance of winning the point. In 2014, she won 68.8% of serves out wide, 65.1% down the T, but just 51.7% of serves down the middle in the deuce court. She has clearly observed this and decided that risking bigger serves down the T and out wide is worth the extra missed first serves as a result.

This is also consistent with a desire to limit the length of points. In 2015, with serves out wide on the deuce court, she has won 46.5% in three shots or fewer, 61.9% of those serves down the T and just 12.4% of those serves down the middle. She is taking more risks with the direction of her serve to shorten points.

We can see a similar drop in serves down the middle on the ad side, with the most noticeable change here being an huge increase in second serves down the T, increasing from 19.8% to 37.3% this year. The decrease in wide serves here is particularly interesting given that it has been a very effective serve when she has hit it this year.

Return Game

Let us now move on and look at her return game in more detail.

Less of a dramatic change here, but we can see an increase in unforced errors in all three situations, combined with an increase in rally winners in two of the three situations. Maybe this suggests that Serena is going for more aggressive shots this year in an attempt to keep points shorter?

There is also a theme in her return stats to suggest that she is attacking more on the forehand return.

We can see that she is getting fewer returns in play, but we can also see a 1.8% increase in the percentage of return winners that she is hitting and a marked decrease in the average number of shots in the rally when she gets the chance to make forehand returns. Again, she is gambling on being more aggressive, but it is not necessarily clear here whether this tactic is working. It could well be that she feels her stamina and her rally game as a whole is declining, so by keeping points short, she is giving herself a greater chance of winning matches, even if she is winning slightly fewer points year-on-year.

We can see here that she is particularly attacking the wide serves. While we can see the percentage of winners increasing on all types of serve, the increase from 3.0% to 8.3% on wide serves is a huge increase and suggests that she is attacking on these returns rather than risking being caught out of position and having to move quickly across the court to recover her position.

Conclusion

As mentioned earlier, I stated last year that I felt Serena was in decline. With another year having passing, this picture is not quite as clear. She is, or believes herself to be, clearly declining in certain aspects of her game. Age is catching up with her and her speed and stamina around the court is not what it once was. This is something that she cannot really prevent. There have also been suggestions of a knee issue, which would also feed into this physical decline.

However, she is clearly adapting her game to overcome these physical restrictions. Her first serve continues to become more of a weapon and she is looking to shorten points by hitting winners whenever she has the opportunity. She is also becoming more selective in the return games that she expends energy in.

Whether all of this means that Serena is improving or declining is unclear. She has undoubtedly had an outstanding 2015 and she continues to retain that aura of invincibility when she steps on court, which gives her a huge mental edge. So many players have had the opportunity to win sets or matches against Serena this year, but few truly believe that they can actually do it, even when only points from doing so. While she continues to have this effect over opponents, she will have a big advantage even before the match starts. However, a few losses can easily start to shatter this aura as we have seen with both Roger Federer and Rafael Nadal in recent years.

It will be interesting to see what 2016 holds for Serena, but it would seem likely that the transition to Serena 2.0 will continue.