Stats and Projections for Spurs-Hull City

Tottenham Hotspur welcome Hull City the Hull Marketing Approved Megafauna to White Hart Lane today. Hull have performed a lot better in the Premier League than I expected them to based on their entirely blah promotion campaign from the Championship in which they only outscored their opponents by a total of nine goals over 46 matches.

By the underlying stats, the Marketing Approved Megafauna have once again been moderately fortunate with their results. Hull have put 12 shots on target from inside the box, 8 from outside, creating 6 big chances and winning a pair of penalties. They have allowed 22 SiBoT / 13 SoBoT as well as 9 big chances and one penalty. The rate of big chances conceded per match is pretty good for a promoted side, suggesting a solid defensive presence. But I have Hull expected to a goal difference of about -4 compared to their actual -2, neither of which quite supports a 3-2-3 record.

Again, they're not bad. By the numbers, this is a capable defensive side that is reasonably difficult to break down. But we shouldn't look at Hull City's 11 points and think that dropped points are an acceptable outcome. These are matches Spurs can't afford to drop points from, especially after already taking a home loss against a better (but still not that good) West Ham United side.

Projections for the Game and the Season

• Expected Goals methodology
• Season Simulation methodology

There's not a lot to be gained from winning this game for Spurs, it's just a matter of not taking a hit for losing it. I'm guessing that the (small) divergence between my projections and the bookie projections has to do with the factors noted above. If Hull City are a somewhat more capable team than I have them projected, if their second place finish in the championship and 11 points this year reflect a quality that I'm not capturing in the underlying stats, then they should stand a slightly better chance of picking up points. A two-in-three chance of winning is still pretty darn high, though, so the bookies aren't exactly blinded by the top line numbers either.

The figures below were initally totally wrong due to a silly error. They are fixed and are now correct.

I wrote a script this week to log every scoreline in the projected games. I'm not honestly sure if that's interesting at all, but I have all these numbers now, you know? My projections have the likelihood of a Spurs clean sheet at 56%. That seems like a mighty high number. Other numbers from the score projection:

1-0 and 2-0 were the most likely scorelines by far at 16% and 15%. Next most likely was 3-0 (9%). Hull scored one goal or fewer in 88% of the simulations. The most likely winning Hull score was 0-1 at 3%, less likely than Spurs 4-0 or 4-1 (5% and 4%). I ran 500,000 simulations of the match and the season, which means there were some super odd outcomes. A (half) million monkeys at a (half) million typewriters and all that. In one iteration, Hull won 8-5, in another Spurs won 9-7. I just thought that was sort of funny.

If there is information from simulated games and seasons that you're interested in, let me know, and I can try to save it for next week and present that data. (Note that my simulations work only at the team level projecting team goals per match, they include nothing to do with individual players.)