Looking forward: a closer look at the Manchester teams

The situation in both Manchester teams has been one of the most important issues so far this year in fantasy, and to a degree, those who have simply ignored the drama and gone with elite players from the London clubs have been rewarded. Of course, Robin van Persie and Carlos Tevez have both enjoyed plenty of success to date, but that has been achieved amidst a shroud of uncertainty, which at least in Tevez’s case is still far from cleared up.

The prompt for this post is United’s prospects over the next eight gameweeks, which under the new model look simply outstanding, particularly at the attacking end of the field

Powered by Tableau

First, as always, a couple of caveats on this data. If we look at what Man United have done so far versus league average, their shot generation statistics are excellent, though not to a degree that would arouse suspicion about their validity. At home they’re generating 40% more shots inside the box (SiB) than the average team has against their opponents, and creating 29% more chances. Those numbers are high but not ridiculous when you consider that Everton (39%, 24%) are very close and a handful of other teams are within shouting distance. On the road they’ve been less dominant with 17% more SiB than the league average which significantly trails teams like Man City (59%), Everton (44%) and Arsenal (35%). In short, while United have shown very strong numbers, they are a very strong team and I’m fairly comfortable suggesting these numbers are sustainable going forward (though I can’t be sure without some research as to how much they could fluctuate over a season).

The next point is one in which I’m less confident and that’s United’s conversion rate of shots into goals. So far this year 24% of their SiB have been converted into goals. The good news (from a sustainability perspective) is that Chelsea aren’t very far behind on 23%, but after that the numbers drop off quickly on the way to the league average of 15%. Are we confident United will remain above league average for the year? Absolutely. In van Persie and Rooney they have two of the best forwards in the league, and with plenty of talent around them, one can make a strong argument that the quality of those shots inside the box will surely be higher and lead to more goals than, say, a team who plays a more direct game and needs to rely on difficult headers from lesser players. The issue then is how high above average United can sustain that conversion at.

Last season United led the league in SiB conversion rate with 18% and while it's arguable that van Persie makes them a better team, I'm not sure than a lack of clinical finishing from the likes of Hernandez inside the box was the issue last year. Therefore, I think it's safe to say that we can expect their conversion rate to fall a bit and thus the below stats are probably slightly overvalued. The good news is that by the time we get to that troubling Sunderland forecast in GW17 (4.56 goals), the data should have levelled off a bit to their actual talent level. With that caveat in mind then, let's look at what the model has in store for the key players from United, and also their city rivals.

Powered by Tableau

This data thrills and terrifies me in equal measure. I came into this piece expecting to bump up my van Persie rating, possibly enough to suggest buying him, though still with some concern over that price tag. Then the data comes out and Rooney actually edges ahead in terms of projected points and of course then comes with a substantial lower price tag. I want to a full write up on the player forecast data but this pair (and we can throw Aguero/Tevez in there too) gives us a good case study to highlight what attributes are driving success in the model, and whether they indicate something that might need to be tweaked.

Red Devil
Let’s start with the Red pair. On face value, van Persie leads his teammate in all significant categories, including of course, a sizeable lead in goals (7 to 2). The shots data suggests it should be closer than that but still that van Persie has an edge. What happens though if we only look at games the two have played together? Van Persie’s biggest game by far was his stellar performance against Southampton is which he racked up over half his shots for the year and three of his goals. That, however, was a game in which Rooney did not feature. They’ve only played six times together so a small sample size warning applies here, but if we go with it for a second, we see that the goal split is now down to 3:2 in van Persie’s favour while the shots are now flipped to support Rooney as the more promising prospect (19 total shots to 12, 9 SoT to 6).

A further, somewhat bizarre twist when you consider the narrative, is that over those same games, van Persie has created more chances (12 to 8) so the common perception of Rooney playing too deep to score but possibly benefiting from increased assists is false so far this season. A second peculiarity to note is that Rooney’s forecast data is more inconsistent week to week than van Persie, leading to the Dutchman getting seven ‘captain picks’ (crudely defined as a forecast over five points) to Rooney’s four. This is mainly driven by Rooney’s struggles away from home which have seen him account for just 25%, 13% and 11% of United’s shots inside the box, compared to totals of 27%, 20% and 25% at home. Again, these sample sizes are tiny and the model shouldn’t be used on that granular level but it’s something to keep an eye on if Rooney continues to be less effective on the road and thus becomes much harder to captain.

When you’re talking about such small samples it’s important not to get carried away and it’s probably not totally fair to simply throw out van Persie’s best game to prove a point. That said, even with that data included I’m struggling to see a ton of evidence to suggest the 1.6m extra outlay is worth it. If Rooney continues to ‘struggle’ on the road to the point he can’t be captained with confidence against anyone but the weakest opponents, we might need to revisit this data, but for now I’m backtracking on my previous hesitancy towards Rooney and recommend him among his United peers.

Elsewhere at Old Trafford the other options are almost impossible to own right now (ignoring the defense that is, which just about remains ownable). Kagawa has the best data which backs up the direction I would go if forced but he’s in a bracket alongside the likes of Fellaini, Pienaar and Morrison who all come with a lower price tag and more certainty in the playing time department. The upside of United’s fixtures is sufficient to make Kagawa a consideration but it’ll only take one or two benchings to ruin those great fixtures and I don’t see enough upside to justify the risk here.

Blue Moon
United’s data is easier to analyze as while they have strength and depth in numbers across the park, Ferguson has shown a tendency to play his top players consistently and thus you can live with the odd rest whenever it comes up. Mancini, of course, is a different animal. Thus, when we look at this kind of model – which totally ignores the prospect of future playing time – we need to always be mindful of a player’s future opportunities (or lack thereof). Once again, for better or worse the model goes against my gut instinct, which was to suggest that while Aguero is probably the better prospect, it was sufficiently close to make Tevez the clear man to own here; now I’m not so sure. The issue is that while Aguero has accounted for 36% of City’s SiB at home and 34% on the road, Tevez has been solid at home 25% but fairly redundant away from Maine Road/City of Manchester/Etihad (delete as per your age) with just 9% of City's SiB. With the requisite small sample size warning, if we look at the two away games they've featured in together we see Aguero dominate against Fulham in GW6 (5 total shots, all inside the box) while Tevez struggled (3 shots outside the box only) while Aguero was also able to also match Tevez's totals against West Brom, despite playing just over half an hour.

Though Mancini appears to have kissed and made up with Tevez and Aguero has felt the cold touch of the boss' rotation policy this year, over the course of a season I'd still back Aguero to get more games. Whether or not some of those are vultured by the Champions League is another issue (nicely discussed by Shots on Target here) though that could become a moot point if City were to slip again this week against Ajax. The playing time is tough to forecast and I don't want to get into guessing games, but unless you believe Tevez has a sizeable advantage in that area, Aguero looks like the better player here, and probably by enough to justify the extra cost (especially if you hold Tevez at 9.5m rather than 9.2m). The fact that Aguero, arguably the best forward in the league is held by just 7% of managers is also a nice bonus.

Elsewhere in City's squad, it's a familiar story than above with their cross town rivals. Silva is just about hanging on to my shortlist, thanks in part to the fact that his away stats are great and there's no real reason why his home stats shouldn't catch up (he isn't after all a Walcott type whose game can sometimes suffer when teams drop very deep). The problem for Silva is that Bale and Mata are both quite clearly better at the moment in terms of the model and also job security and with the aforementioned forwards looking very promising  it's just tough to fit Silva in right now. The model likes him more than that mid-level group of Fellaini, Pienaar and Nolan but I'm sure it's by enough to warrant the extra cost.

I'll try to not make this personal but after delaying signing Mata due to his fixtures and grabbing Yaya who had much easier games, I'm a bit bitter. Anyway, the model doesn't like Yaya much more than I do right now, due to measly 6% SiB share at home and 10% SiB on the road. His assist numbers are a touch better (14% and 17% share of chances created) but he's consistently being held back and no longer appears to be 'let loose' towards the end of games as was the case numerous times last year. Nasri doesn't fair much better in the model and has the added knock of not being guaranteed playing time. I don't see either of this pair as very good options right now and the 33% of us who own Yaya should really be targeting him as a player to move sooner rather than later, even though City's fixtures are pretty good.

So we're left with a tale of two pairs of forwards (and I guess Silva). At United I'll lean towards Rooney and the extra money where as at City I'll pay the premium to grab Aguero over Tevez. The added bonus with those moves is that Rooney (8%) and Aguero (7%) are held by significantly less managers than their teammates (van Persie 32%, Tevez 42%). Remember of course that the model is still a work in progress so there's still plenty of other considerations to take into account, but I believe in the data and I'd be comfortable making these moves in my own team.