Player Analysis: By The Numbers Week 17
There is a dark-side to probability, a side that no one likes to talk about. It lingers in the background, out of sight and generally invisible to football fans. I like to call it “the fickle and angry god of 100%.” You see, when you predict the scores of matches you display them as a score out of 100% with the only two absolutes being 0% and 100%. That means that there are only two cast iron guarantees and neither of these percentages ever actually show up.
After Spurs lost to Newcastle (I know, I’m a spurs fan) I received a twitter message from a reader @soccerdfs simply saying ‘81% eh lol’. I sympathised. The highest score to have come out of the predictor in a while had got it wrong. Or had it? You see, the dark side of probability is that it is proven over the long term, not over one game. Let’s say that we have 100 cards, 81 of them have Spurs on the back, 11 have Newcastle on the back and 9 have draw written on the back. If we could pick a card 100 times then we are guaranteed to finish with the percentages that the predictor specified, if we can pick only one then we are more likely to pick Spurs but certainly not guaranteed. We could however unlikely pick 19 cards in a row and still not manage to pick Spurs. You see, we are dealing with likelihoods, not dead certs.
Another way to think about it, and the way that keeps me sane when the most unlikely results come in, is to imagine that you’re Bill Murray in Groundhog Day. As every game would be played in the same weather conditions, with the same fans and the same players we’d have the chance to see how things would happen when repeated numerous times. Whenever you see a percentage probability like last week’s you should always think to yourself ‘Newcastle have an 11% chance of winning so if we could play this exact game over and over again, they’d win 11 out of every hundred’. Football prediction is a question of putting your best foot forward but understanding that even the most unlikely events sometimes occur. That leads me on to…
Could this finally be the week when Chelsea get a win? I hope so for their sake. I have looked through my correct score prediction spreadsheet which isn’t usually included on these pages and it predicts that there’s a 78% probability of one of the following scorelines: 1-0, 3-0, 4-0, 4-1, 2-1, 2-0, 3-1 or 5-0 – none of which sounds particularly good for Sunderland. Although Chelsea are struggling on all fronts so who knows? As my brother text me the other day “I woke up to shouting outside my window yesterday morning, looked out and saw four blokes in Chelsea shirts playing football with a kitten. Shocked, I was just about to call the RSPCA when the kitten went 1-0 up.” On that note, let’s look at the predictor.
Predictor
The predictor is a subtle affair this week with the highest probability being 70% (remember 3 times out of 10 it won’t be a winner). The teams you should be fancying are United, Newcastle (really), Chelsea (again, really), Arsenal and West Brom. These teams have six game goal differences that absolutely dwarf their oppositions massive negative performances. I’d be looking for Mata, Costa, Ozil, Ramsey and Rondon. Except for Arsenal, It doesn’t sound like a convincing group given recent performances but their opposition have conceded a staggering 39 goals more than they’ve scored over their last six away games.
You’ll note that Spurs move from last week’s favourites to this week’s 2nd away favourites, a drop of over 50% probability. This is because there is a massive advantage to the home team and don’t ever forget that. It is also highlighted by Leicester, who besides having one of the highest away tallies I can remember with 8 still have a lower probability than Everton or the draw.
In previous seasons, I would have said that you should be loading up on Chelsea as they have the most obvious match up. I can’t however in good conscience even though I’d love to. I might bring in Courtouis as they’ve got four easy games coming up against Sunderland, Watford, Palace, West Ham and two trickier ones against Everton and United but I’m really not convinced. De Gea, squaring off against Norwich, Stoke, Chelsea, Swansea, Norwich and Liverpool looks a safer bet.
The Poisson
The Poisson quite rightly suggests that the game most likely to have goals (and pretty much a banker for readers who like both teams to score bets) is Everton vs Leicester. Those of you with Lukaku and Vardy should hold on tight. The polar opposite is the Watford vs Liverpool game that shows nearly a 70% chance of under 2.5 goals. I’m not convinced with Liverpool’s forwards at the moment, Fimino in particular but I am convinced with Ighalo and I won’t be dropping him anytime soon.
The Poisson suggests avoiding Everton, Leicester and Spurs in your defence – I’m not convinced by Southampton lately so I’m not sure about the last choice but maybe that’s my allegiance coming through. It also suggests that you should focus on Newcastle, Crystal Palace and Liverpool for your defences but I can’t see any premier league player figuring out who to sell to bring in Coloccini or Mbemba. Sometimes I have to just put my hands up, explain why the maths points to certain figures and leave you guys to draw your own conclusions. See you next week.
This entry passed through the Full-Text RSS service – if this is your content and you’re reading it on someone else’s site, please read the FAQ at fivefilters.org/content-only/faq.php#publishers.
