That's mainly because I don't track data that intensively, it's just not a market that I'm interested in - I didn't realise that, until doing some basic research, the general market is actually three way with a "can bet the tie" option, I initially thought that just on brief sightings when I'm scanning through the markets for comedy value that they typical market was two way and that a tie would be regarded as a push. I also think it's too easy a market to manipulate - while anyone that pays the slightest attention knows that, say, Madars Razma goes for T19 a disproportioniate amount of the time, it would be extremely easy for someone who wanted to throw the market to switch at what looks like a perfectly reasonable time for strategic reasons. Start with back to back ton 40's to be on 221 and then start with two trebles on visit three with your opponent in a spot where they can leave a likely checkout? Going down to bull to leave 76/51 and not 81/41 looks extremely reasonable to me, and that's not taking into account of many a time where I see players (notably the Dutch) have a mediocre first visit then go down to 19's for no obvious reason whatsoever.
So what caused me to make this post was when I was looking at the Slam last sixteen around this time last week. I had Edgar's watch along in the background. Now Matt's a pretty knowledgeable guy, and does like his stats. I'm not sure that he's qualified as I am to talk about it, but he's at least done one video where he's done a breakdown in terms of a checkout (I think it may just have been a redo of the Deller 138, but I could be mistaken and it could have been more recent) so I'd say he's probably more comfortable with the numbers than maybe 80-90% of players, and especially pundits just based on that one video. What really surprised me was when he was talking about the Ross Smith against Martin Lukeman game. Now here, he'd produced some stats in terms of how often the two have hit 180's. Everyone that's paying attention knows that Smith is a huge 180 hitter. In terms of Lukeman, I can think of things to describe him - good finisher? Sure. A bit Wadey in terms of taking chances? Yeah, sure, I think I've used that analogy in relation to Smash before. A power scorer, maybe not so much. What Edgar did produce is that Lukeman was hitting a 180 in every 0.18 legs, while Smith was getting it in 0.37 legs - at least on seasonal data. Where he's derived that from, I don't know, but I'll take him at face value. That is twice more likely per leg for Smith. What was disappointing is, and I quote, Edgar saying "there's no value in 180 betting whatsoever" when looking at the prices and it seeing 1/8 on Smith. That might be true, but that's lazy.
To me, my initial thought is probably "he might not be wrong, but being twice as likely to hit a 180 in a leg over a match that is going to run for a minimum of 10 legs, and maybe up to 19, where your chance of hitting a 180 is more than one in three (I mean if you chuck two bad pub players against each other and say first to ten then the 180 race probably ends 0-0 but these players are a tad better than that)? I'm not sure that we can be quite so dismissive and say "lots to one on clearly no value lel". But how do we model this? Now here, unfortunately we do not have the greatest data available - namely because there are instances where a player can hit two 180's in a leg. Clearly, the most obvious situation is where someone gets on a nine, but we've all seen plenty of cases where someone's gone 180-60-180 or hit back to back after a shit first visit. That's fine. What I think we ned to look at is, at least in the pro game, how many visits a player will get where someone can hit a 180.
If we think about this at the most basic level, you can score up to 319 points before you can't score a 180. That's fundamental. Once you're down to 181 required, you're fucked, you're not scoring any more. So at a pro level, you're typically saying you will get at most four visits where you can hit a 180. If you're averaging less than 80 when going for pure scoring, that's not good, sure there will be the odd leg where there is a serious scoring malfunction and you get one or more visits after that, but those are outliers. The point of what I'm saying is that it'll take a lot from the opponent to limit that four to a smaller number - he's either going to have to start on throw and hold in four himself, or hit a nine. I think it's not unreasonable to say your chances of hitting a 180 in a leg are independent of what your opponent is doing. Not ideal, but let's roll with it.
Similarly, to go back to the chances of you hitting two 180's? I think that's also moderately infrequent. You can only score 320 points before you can't hit a 180. If you know you are going to hit a 180, that only leaves 140 points to play with - score that or more, and you're done. Open 140? You're fucked. This isn't something we can go into with any sort of Poisson distribution analysis, as the probabilties will change once we've hit one. Hitting a second 180 in a leg is a lot, lot harder once you've hit a first one, if you're even able to hit a second at all. As such, I think it's also not unreasonable to just model on pure 180s/leg stats and ignore the possibility of hitting two in a leg - again, not ideal, but if we factor in times where we'll hit two and the other guy will hit one (or two for that matter) back, it gets drawn back a tad. We just need to be aware that whoever's projecting to hit more 180's per leg may be slightly underestimated in terms of how much of a favourite they are.
So let's crunch some numbers. Edgar's saying Smith hits one every 0.37 legs, Lukeman every 0.18 legs. If we're ignoring the small possibilities where someone hits two, we basically have four outcomes:
- both players hit one
- neither player hits one
- one player hits one and the other doesn't (each way)
So in the case of Smith against Lukeman, this says that Ross will gain a 180 point (for want of a better term) 30.3% of the time, and Martin will gain one 11.3% of the time in any given leg. All we need to do is to expand this out across the match.
What does this give us? Well if either player won 10-0, Smith would get the most 180s 77% of the time. Heck, if it went the distance Smith only gets up to 88% of the time, which is break even for the price that Edgar was quoting. Lukeman was only at 11% on a 10 leg match, and down to 6% on a 10-9. As it turned out? We'd have been "don't bet either lel", Lukeman would sneak home 6-5, and Edgar would technically have been right. Thanks for wasting my time Matt!
No comments:
Post a Comment