Scorecast Stats Bulletin - issue #1

Introduction

Greetings, Scorecasters and other interested parties. This is the first of many blogs I'll be doing, taking the predictions of the Top 10 Scorecasters in the league and seeing if they might inform the weekend's betting, for those who are that way inclined. This first one will probably be longer than most as (a) I'm not sure what information is going to useful and what isn't, and (b) I'm going to go into a bit of explanation as well. The explanation is in "The Blurb", just below; if you want to skip that and just read "The Actual Information", you can click "The Actual Information", or just scroll down a bit till you see the headline.

The Blurb

So... for those of you who don't know, don't care, or haven't been paying attention... here's how this is going to work.

Throughout the season, there are 24 rounds of the Scorecast competition and there is now a total of 106 entrants, including four new entries at Round 2. I hope and expect that this number will grow. With this being the case, we've actually got a pretty decent sample size for stats as well as the top-10 to compare it to. I’m hoping that might just be enough to make some educated guesses on the football scores and results, for each Scorecast round.

This bulletin, given that it might give an unfair advantage to people who’ve not submitted their scores, is being sent by email but after the actual deadline (12:30 tomorrow, Saturday) it’ll go up on the website as a blog post, so people can refer back to it later, or comment on how they fared, if they have indeed put on a bet or two. Within the last 24 hours of the deadline, it will be impossible to overwrite your scores using the app (it’ll tell you to email me instead if there’s an obvious typo, but it wouldn’t be fair for people to be able to change their scores after they’ve received this document - see pic). Therefore, there are now two deadlines per round: The Overwrite Deadline, which is 24 hours before kickoff of the first selected game, and the Scorecast Deadline which is exactly upon kickoff of the first selected game. As such, people who’ve not submitted their scores for the round by the Overwrite Deadline will still be able to submit their scores within those last 24 hours, having not received the Stats Bulletin.

The Stats Bulletin might not be every week (it’s fun experiment really, so I’m not prioritising it) and, in fact, it might not go out bang-on the Overwrite Deadline time (QED) as I tend to have a fair bit of coding admin to do within those last 24 hours so I’m often a bit pressed for time. If it goes out late (like today) it will go out to everyone who’s done the Scorecast by that point, regardless of whether they were in before the Overwrite Deadline. That does mean, though, that if you complete your scores after the Overwrite Deadline, chances are you won’t get the bulletin unless you ask for it, which you can do by emailing rob@toontalk.uk

The Stats Bulletin will not necessarily be the same every round. I’ve only tinkered with one or two metrics so far, and I hope to add more to that (in fact, I’m very much open to suggestions). So, you might get some new stuff, next round… or you might not. We’ll see. Anyway, that’s enough of a blurb and I’m sure most of you didn’t read it anyway.

Without further ado, let’s move on to…

The Actual Information

Firstly, I’m going to run through the Round 1 stats, as I shared this with a handful of people by way of a test.

ROUND 1

These were the average scores predicted by last season’s Top 10 players (because obviously there was no Top 10 for this season, before Round 1). Now, I had access to this quite early on (because I compiled it in the first place) and, as such, I put my bets in quite early on.

Thing is, early on, there was at least one of the Top 10 who hadn’t done their scores yet… and if I’d just waited a day, the ONE RESULT that was different would’ve been 1-1, as on the right, not 1-2, as on the left, for Chelsea and Liverpool. The below table was what the consensus of all 96 Scorecasters (that’s how many there were at that stage) thought the scores would be. As such, the bet I put on for six out of eight win/lose/draw matches to come through WOULD HAVE WON. Admittedly, it would only have won me about £30 (I needed all 8 for about £1.6k) but still, that’s absolutely gutting.

This time, seeing as I wasn’t technically incorrect in the way I placed my bet, I’m going to be doing the same thing again. £5 on a 6/8 bet (or £4.76, because it’s 17p x 28 combinations of 6 results, as pictured on the right). This isn’t all of them, but if I get any 6 of this week’s 8 result picks (which I’ll show you in the ROUND 2 section) can win somewhere between £1.90 and just over £10. If I get 7 out of 8, I’ll win somewhere between £25 and £60, and if I get all 8 correct, it’s £185. Incidentally, if anyone wants a good calculator for working out odds on these types of bets (and Goliaths etc. as well, for those of you who remember doing them in the Scorecast), I highly recommend ACE ODDS. Anyway, I won naff all last round because I didn’t wait til more people had put their scores in. If anyone was thinking about waiting til the last minute, it’s pointless this time because the Top 10 have all already submitted, but I hope to encode a dynamic table into future bulletins so that you can check the status quo, as opposed to only checking what the status quo was at the Overwrite Deadline. For what it’s worth, the other “most suggested” bets from the last round were as follows:

That’s a 64% success rate for the bets I picked based on the data. Not terrible. Of course, you’ll see the data from Round 2 below so you can pick your own, but I’ll suggest some anyway. Just for simplicity’s sake though, this was my summary last round, after presenting the data. All of that, on the right, proved to be correct. That’s not a bad record at all and I’ll be EXTREMELY pleased if the below info produces similar results. Now, for the important bit…

ROUND 2

The second round of the competition has seen 4 new entries join us which hopefully means there’ll be a full cohort of 106 for this round. As I write this, we’re at 97 but I think we’ve actually managed a full boat for Round 2 on most seasons… it’s usually after that, that someone or other tends to fall away. To be honest, I’m not expecting a full boat this time… there’s a lot of people, there’s at least one person who’s not got a valid email address… ANYWAY… here are the results as predicted by the current Scorecast Top 10:

There are quite a few gleanable bits of info here, but the main one I’m going to pick from this table is the average scorelines, ie. the average no. of goals predicted per team.

The data says:

Now, personally, I think that looks pretty tasty for a Goliath as we have six 2-1 games (one way or another), a 1-1, and a 3-0. Those three scorelines, ignoring home advantage, account for 32.4% of all football games ever played, according to footystats.org. If we forget the scorelines, just W/D/L on those games all look pretty nailed on as well in my humble opinion… with the possible exception of our own (I do plan to investigate, at some point, based on the data, just how rose-tinted our fan base’s specs are).

Speaking of which, what have the consensus gone for?

Not a colossal amount of difference there - the only scoreline that differs is that the Top 10 reckon City will only win 0-3 at Bramall Lane, rather than 4-0. It’s the similarities that are interesting though, and it’s worth remembering that the consensus does include the Top 10, so anything they’re unanimous on is probably worthy of note. So… I’m not going to explicitly suggest what bets to put on, save the Goliath I mentioned (which is a minimum of £24.70 ie. 0.10 minimum bet multiplied by 247 different possible combinations of 2 or more winners), and the one I’ve put on myself (also mentioned earlier), which is this:

Rather than explicitly suggesting bets, I’m just going to explain what the data suggests might be good, if it is reliable. Bear in mind, I’m good with numbers, but I am NOT a professional statistician (in fact, I got a D for Stats A-Level). It’s also worth noting that, with no disrespect to the current Top 10 (who, by the way, are Yin Mo, George Bland, Steve Beggs, Stephen Ferrell, June Payne, Kjetil Rykkje, Cameron Richardson, Bill Wonders, Jon Gretar Levi Jonsson, and Stephen Hill), it’s hardly a “form table” after only one gameweek. I mean.. I’m 12th out of 102 for cryin’ out loud! 8 of the current Top 10 are new entries, two of them are seeded #35 and #36 based on their performance over the last 2 seasons, and that leaves current champion and overall #2 seed Stephen Hill as really the only member of that group that has any demonstrable merit. That said, Yin and George did both get 8 out of 8 W/D/Ls correct last round which is pretty spectacular. Anyway, the most important thing to remember is that I’m putting this information out for FUN. I’ve already disclaimed til I’m blue in the face, but it’s certainly worth mentioning again that neither Toontalk, nor the online entity that is Roberto Di Statteo, nor the real human person that is Rob Waters, nor any of the other Toontalk Admins, accept any responsibility for anyone’s gambling losses. We do not have a sure-fire prediction algorithm (yet) and we do not claim to have a golden ticket. This is just me presenting the data of your existing predictions and occasionally suggesting a bet based on that but, mostly, just giving you data from which to draw your own conclusions. So, if you lose money betting on it, it’s your own damned fault.

So, without further ado, here are my picks for Round 2:

• Nobody from the Top 10 has picked Bournemouth, Fulham, Crystal Palace, Nottingham Forest, or Sheffield United to win. As such, an acca on them all to lose or, safer, double chance not to win, might be a shout.

• 9 of the Top 10 think that Bournemouth will score exactly one goal. The only one who doesn’t is reigning champion Stephen Hill. 83.5% of the cohort say that Bournemouth will score at least one goal against Spurs.

• 100% of the cohort (and, therefore, the Top 10) have Man City to beat Sheffield United at Bramall Lane.

• Man Utd, Brighton, and Newcastle each have 8 of the Top 10 Scorecasters backing them to win. The figures for the full cohort are 83.5%, 80.4%, and 85.6% respectively.

• Only 7 of the Top 10 have slated Arsenal to beat Fulham; by contrast, 94.8% of the cohort have Arsenal to win. Do the Top 10 know something we don’t?!

• 96.9% of the cohort think Brentford won’t win at Crystal Palace. This is mirrored by 100% of the Top 10.

• 85.6% of the cohort, and 8 of the Top 10 think Man City will keep a clean sheet at Sheffield United. 9 of the Top 10 think Man City will score 3 or more goals, with 89.7% of the cohort thinking they’ll score at least 4.

Make of those what you will, folks. Sorry this one’s so late… and so long. I won’t have to do all the blurb every time, so the next one will be comparatively concise… but then, The Stand is concise compared to War and Peace…