Regular readers will know that this series we have written four articles making the case against certain acts winning the show. All were prompted by acts being heavily backed down towards the head of the market, in what we suspected was an over-reaction to one good performace. Two of the three acts we wrote these articles about after week 1 have gone – The Risk and Craig Colton – while the third, Misha B, clings on by her fingertips.
We probably should have quit while we were ahead, instead of being tempted to give the format another outing when the plunge on Little Mix happened after week 4. In the comments to that Little Mix article, bunnyman said: “The difficulty with any gambling related article which starts with ‘why I don’t think x will win’ is it gives no consideration to the price and whether it offers value.” We thought it might be interesting to write a post on why we didn’t frame the argument this way.
There are a couple of reasons. First, we want to make sure Sofabet is as accessible as possible to readers who enjoy dissecting the machinations of the show, but who are not necessarily seasoned punters. Second, we think there are limits to how far the concept of value betting can be applied, especially in X Factor betting.
Before we explain why, let’s quickly outline what “value betting” means for any readers who may be unfamiliar with the term.
Value betting was popularised in a book of the same name by a tipster called Marc Coton, who sprang to fame in horse racing circles as the Racing Post’s first Pricewise columnist in the 1980s. Essentially, the idea is that you shouldn’t get hung up on whether your individual bets win or lose, but instead on whether they represented “value”.
A coin toss is a classic example. Suppose I offer you this bet: Heads, you give me £10; tails, I give you £11. You accept, we toss the coin, and it’s heads. Bad luck – I win and you lose. But you should nonetheless be happier than me with the bet we struck, because it was good value for you and bad value for me.
Why? Because if we’d split the world into 100 parallel universes at the moment we struck that bet, we’d expect to get something like 50 heads and 50 tails. Whatever the result of any individual coin toss, in the long run a punter who keeps backing tails at odds of 11/10 will do better than one who keeps backing heads at 10/11. The true price was evens.
This is a useful way of thinking about many bets. We can, for example, imagine last Saturday’s 1-1 draw between Manchester United and Newcastle United being played in 100 parallel universes and ask: In how many of those universes would Manchester United have won? Things would have played out differently in each, as countless events – Tim Krul’s fingertips reaching a goalbound ball or not, a debateable penalty being given or not – would have propelled the game into different directions.
Let’s say, having watched the match, we conclude that Manchester United would have won in 80 of those parallel universes. They were available at 2/5 before the game, which equates to winning 72 out of 100. In which case, we can conclude that if you backed Manchester United, although your bet lost, it was good value. They should have been 1/4.
Of course, we can’t experimentally verify our postulation of 80 wins out of 100, as we could by tossing a coin 100 times. But the question makes sense, and it’s a useful way for punters to frame their thinking about many events.
Many, but not all. Imagine, for example, that you are having an argument with a friend in a pub about obscure matter of trivia – say, who sang the opening line in Rhythmix’s performance at judges’ houses. You vaguely recollect it being Perrie, your friend is pretty sure it was Leigh-Anne. Your friend suggests you offer him some odds and then you look it up on YouTube to settle the bet.
If we split into 100 parallel universes at this point, then either you would be right in all 100 or your friend would be right in all 100. In this case, there is an underlying reality of which neither of us are aware, and the concept of value betting ceases to be a useful way of thinking. You just have to go with your gut reaction of how confident you are.
Is betting on the X Factor more like a football match or a pub argument? Consider the week 4 singoff between Misha B and Sophie Habibis. As he wrote in the post-mortem post, Daniel backed Sophie to be eliminated to the tune of around £13,000 at odds averaging 1.08, thus making an 8% return – a profit of around £1,000. TomB asked a pertinent question in the comments: “in my view there was a chance (albeit a small one) of Misha being chucked… so I wouldn’t say 1.08 was necessarily ‘free money’ – do you think it was that far off being priced right?”
For it to be “priced right”, in value betting terms, we would have to say that if that singoff happened 100 times then Misha would be saved in 92 of them and Sophie in 8. But that doesn’t make any sense. There is no randomised process analagous to a coin toss or a goalbound deflection that would determine whether the word “Misha” or “Sophie” comes out of the judges’ mouths at the critical moment. Rather, there is an underlying reality of which we are not aware, which dictates that either Misha is going to be saved in all 100 parallel universes or in none of them.
(Actually, that’s not strictly true. On live television there is always the possibility of a screw-up, such as might or might not have happened in the Katie Waissel vs Treyc Cohen singoff in 2010, when Dermot said he ran out of time to come back to Cheryl and allow her to take it to deadlock. Maybe Sophie might have survived in one).
To take a slightly less extreme example, consider last year’s final. Going into it, Matt’s odds were around 4/5, Rebecca 5/2 and One Direction 10/3. That translates into about a 56% chance of Matt winning, a 28% chance for Rebecca and a 23% chance for One Direction. Yet, if you’d had access to the voting statistics up to that point (which, of course, punters didn’t), you would have seen Matt winning every single week apart from the first one, and One Direction never getting anywhere near the top of the leaderboard.
Even then, things could have gone wrong for Matt in the final – he might have fluffed his lines, or Rebecca might have conjured a moment of magic – but if you ran that final 100 times, how many would Matt have won? A lot more than 56, that’s for certain. About 90, maybe? And how many would One Direction have won? Far fewer than 23, no doubt – it would have taken an act of god. Maybe 1, if they were lucky?
And yet, if Matt Cardle had been trading at 1.1 on Betfair and One Direction had been trading at 100, you would have had punters confidently making the case that Matt was terrible value and One Direction were terrific value. It would have looked that way, but it simply wouldn’t have been so. Though there was no way of knowing it, the underlying reality was such that Matt had a great chance and One Direction had almost no chance.
The current betting suggests that if we split now into 100 parallel universes, Little Mix would win in 40, Marcus Collins in 38, Amelia Lily in 14 and Misha B in 8. When we get the voting figures at the end of the show, we will quite possibly find that one of those acts ought to be at a lot shorter odds right now. Perhaps one of Little Mix or Marcus is way ahead. Perhaps it’s all very tight and producers have something up their sleeves to help Misha, or even Amelia.
Value betting is a extremely tricky concept to apply in such situations. The word “value” becomes instead a kind of shorthand for how confident you are that you’ve correctly guessed what the underlying reality looks like, in relation to the risks and rewards on offer. This is a whole lot more abstract, ephemeral and subjective a calculation than how many times out of 100 a particular eventuality might happen.
So when bunnyman’s comment to our Little Mix article continues: “Do I think Little Mix will win? Maybe, maybe not. Do I think they win at least 1 time in 12 to create value at the current odds? Absolutely”, we just don’t find it useful to think in such terms. Probably the underlying reality at that time (their vote performance up till then, producers’ intentions and so on) was such that they’d either have a much, much better chance than 1 in 12 or a much, much worse chance.
Fortunately, punters always have the option of sitting it out, and it’s one we take most of the time. You may have noticed that while we love theorising, speculating and debating here on Sofabet, it’s relatively rare for us actually to flag up a bet as one we’re having ourselves.
So with Little Mix, for example, we weren’t remotely tempted to back them at 12/1 (oops). But nor were we anywhere near confident enough to lay them at 1/12. Daniel wrote last year about how he woke up in cold sweats having laid Wagner for £28,000 at similar odds in the win market, wondering if he might have completely misread the situation. There aren’t many 1/12 shots where you can be Sophie Habibis confident – and even then, there is always the chance of it going horribly wrong.
How do you conceive of your own punting strategy, in terms of when you decide to get involved or stay on the sidelines? Do let us know in the comments box below.
[UPDATE, 30/11: This article sparked a good deal of confusion in the comments below among seasoned punters who adhere to the “value betting” approach; if you are one of them and your immediate reaction is also to think we have taken leave of our senses, we highly recommend you skip to commenter patrickstar’s very helpful reframing of the above discussion in terms of epistemic versus aleatoric uncertainty, here and here. We are obliged and edified now to be able to use this new vocabulary.
To sum up: Our intention was not to suggest that “value” becomes an entirely meaningless concept when there is an unknown underlying reality (or an “epistemic uncerainty”), and that punters stop looking for “value”, as several commenters thought we were saying; but rather to observe that while one can be objectively confident of what represents “value” in purely “aleatoric” situations, such as a coin toss, this becomes less and less possible as uncertainty becomes more “epistemic”.
We personally find that it clarifies our own thought processes to conceive of what it means to look for a “value” bet slightly differently when uncertainty moves away from being purely “aleatoric”. If you find this line of thought interesting, you may enjoy reading the debate that ensued in the comments thread below.]