How Accurate are the Markets....?

[salmanim140i]

Hey All,

Was reading a good blog the other day and it was talking about the Mass Psychology in Betting (Sports Trading)

It was putting across that 80%+ of the time the markets are accurate in regards to knowing how a Game/Race?Match is likely to finish. This I guess is then reflected in the odds and the percentage this tells us of how likely the end result is going to be.

I get how looking at the odds can determine what the likely outcome will be but if you look at Horse Racing for example the Favorite I believe only wins around 33% of the time even though the odds may reflect that it has a 50%+ chance of winning. Again with Tennis or Cricket the odds are constantly moving up or down depending on who is Serving/Bating and of course who is scoring points.

I guess what I am trying to say is how Accurate is this... Should this be something to take into consideration when trading (betting). I know there is the over round and Peter Webb has done a good few Videos that explain how "efficient" the markets are. But is that just to reflect how good it is to use an exchange than a bookie as this will obviously increase value in your Trade/Bet and cut out any Bookie Profits?

Just trying to Gage how good it would be to use the market as a guide to trade or back/Lay something along with any pre analysis (I do think sometimes using any pre Analysis can be dangerous as you tend to interpret the data for what you want) but if the market is a good guide to go by you could use that to determine the likelihood of a result right?

Just a thought and a ramble but be good to see what others think?

Thanks

[Derek27]

How accurate markets are depend on the nature of the event. Premiership football, Grand Slam tennis, races with top-class horses or seasoned handicappers can have odds very close to their true chances whereas a 15-runner two-year-old maiden where little is known about the ability of the runners are bound to be best guesses. Ultimately, when it comes to punting, the winners will be the ones who can judge probabilities more accurately than average and take advantage of the small price fluctuations in the competitive markets and the larger fluctuations in the more volatile markets.

[ShaunWhite]

but if you look at Horse Racing for example the Favorite I believe only wins around 33% of the time even though the odds may reflect that it has a 50%+ chance of winning.

You're confusing the likelihood of one specific horse winning with the likelihood of any favourite winning. There might be 1 in favs that win but they'll be a mixture of horses at various prices. Favs at 2.0 are going to win about half the time and those priced at 4.0 will win about 25% of the time.

The idea that the market will be 'right' 80% of the time is nonsense. They'll vary from being exactly right thru to completely wrong on a sliding scale. But even then that's only visible when to look at a large number of results, when a horse or team wins or loses at 2.0, it's not possible to know if that individual horse or team should have been 1.7 or 2.7 or any other price. All you know is that over several hundered priced at 2.0 approx half of them won.

[salmanim140i]

Two Interesting views there.

First one is indicating the more volume (people betting) in the market can get close to true probability of an event out come.

The second one there from your self Shaun is indicating that the chances of something winning at that said Probability is the likely hood over hundreds if not a few thousand attempts.

I suppose both are right.. but makes me think how do these Bookies work out what odds to give and then add their commission into it so that they can achieve a profit.

If its based on volume of matches or races then surely on the day that does not stand for anything as anyone is likely to win... but if its based on the first instance then I can see how the Market can come to a general consensus of who or what will win as the theory of the mass market will then have a good understanding of the likely outcome??

Or Have I totally misinterpreted both views?

[Derek27]

Two Interesting views there.

First one is indicating the more volume (people betting) in the market can get close to true probability of an event out come.

The second one there from your self Shaun is indicating that the chances of something winning at that said Probability is the likely hood over hundreds if not a few thousand attempts.

I suppose both are right.. but makes me think how do these Bookies work out what odds to give and then add their commission into it so that they can achieve a profit.

If its based on volume of matches or races then surely on the day that does not stand for anything as anyone is likely to win... but if its based on the first instance then I can see how the Market can come to a general consensus of who or what will win as the theory of the mass market will then have a good understanding of the likely outcome??

Or Have I totally misinterpreted both views?

It's not the number of people betting but the skill of the people and amount of information available concerning the event.

How do what bookies work out the odds to give? If you're talking about layers on the exchange, they don't. They use their judgement which is just as fallible as the judgement of punters. There is skill on both sides of the book.

As Shaun alluded to, betting has nothing to do with what will (likely) win, but matching probabilities against odds on offer, to distingush the over-backed selections from the overlooked.

[ShaunWhite]

Re market accuracy, forget maths let's just use logic. How do you work out the exact chances of something winning on a given day? You can't, And that's why the highest rated horse in the race won't be 1.01 even though it's the best one, so should win. The 'market' will be saying something like 'this horse should win half the time so it's priced at 2.0'. On that day it might be a 1.01 dead cert if all the others have been sick but on another day it's true price might be nearer 10.0 because it's just not in the mood for it. That's where the pro value punters make their money, comparing the prices on offer with their estimate of the outcome on that day.

Bookies don't generally need to get early prices very close because they'll only let you have a tenner on at those prices. Once the markets get into full swing and prices are being set by the hive mind, then they'll just do a basic accountancy job controlling their estimated sp liability by turning the taps on and off on certain horses with their prices. If they've had enough money on the fav they can slash its price, because they don't accept lay bets, that stops the money coming over the counter for it and income starts to come on other horses. Trackside bookies aren't much different, they're unlikely to see enough business to create nice balanced books so almost everything they do is offered a bit worse than the exchange and the apprentice just lays it off (backs it) on the machine. That's why I think prices often spike to round numbers, someone's just getting things hedged and doesn't want to risk being 'clever' drip feeding bets in. If they take a bet for 2 grand at evens the lad will take it all down to evens if necessary rather than messing about and getting caught out.

As Derek says wisdom comes into it, a £200 dog race where just the fanatics are involved might produce a more accurate price than England vs Germany where huge numbers are making partisan judgements based on nothing but hope or loyalty.

[Fugazi]

Re market accuracy, forget maths let's just use logic. How do you work out the exact chances of something winning on a given day? You can't, And that's why the highest rated horse in the race won't be 1.01 even though it's the best one, so should win. The 'market' will be saying something like 'this horse should win half the time so it's priced at 2.0'. On that day it might be a 1.01 dead cert if all the others have been sick but on another day it's true price might be nearer 10.0 because it's just not in the mood for it. That's where the pro value punters make their money, comparing the prices on offer with their estimate of the outcome on that day.

Bookies don't generally need to get early prices very close because they'll only let you have a tenner on at those prices. Once the markets get into full swing and prices are being set by the hive mind, then they'll just do a basic accountancy job controlling their estimated sp liability by turning the taps on and off on certain horses with their prices. If they've had enough money on the fav they can slash its price, because they don't accept lay bets, that stops the money coming over the counter for it and income starts to come on other horses. Trackside bookies aren't much different, they're unlikely to see enough business to create nice balanced books so almost everything they do is offered a bit worse than the exchange and the apprentice just lays it off (backs it) on the machine. That's why I think prices often spike to round numbers, someone's just getting things hedged and doesn't want to risk being 'clever' drip feeding bets in. If they take a bet for 2 grand at evens the lad will take it all down to evens if necessary rather than messing about and getting caught out.

As Derek says wisdom comes into it, a £200 dog race where just the fanatics are involved might produce a more accurate price than England vs Germany where huge numbers are making partisan judgements based on nothing but hope or loyalty.

Been thinking about this. On the flipside, theres no incentive for a team of people to analyse and work on a prediction model for greyhounds because their costs cant be met with the low liquidity.

However, a team of 10 data scientists making a prediction model for England Germany where they could maybe get £500k of bets on...

[Anbell]

Re market accuracy, forget maths let's just use logic. How do you work out the exact chances of something winning on a given day? You can't, And that's why the highest rated horse in the race won't be 1.01 even though it's the best one, so should win. The 'market' will be saying something like 'this horse should win half the time so it's priced at 2.0'. On that day it might be a 1.01 dead cert if all the others have been sick but on another day it's true price might be nearer 10.0 because it's just not in the mood for it. That's where the pro value punters make their money, comparing the prices on offer with their estimate of the outcome on that day.

Bookies don't generally need to get early prices very close because they'll only let you have a tenner on at those prices. Once the markets get into full swing and prices are being set by the hive mind, then they'll just do a basic accountancy job controlling their estimated sp liability by turning the taps on and off on certain horses with their prices. If they've had enough money on the fav they can slash its price, because they don't accept lay bets, that stops the money coming over the counter for it and income starts to come on other horses. Trackside bookies aren't much different, they're unlikely to see enough business to create nice balanced books so almost everything they do is offered a bit worse than the exchange and the apprentice just lays it off (backs it) on the machine. That's why I think prices often spike to round numbers, someone's just getting things hedged and doesn't want to risk being 'clever' drip feeding bets in. If they take a bet for 2 grand at evens the lad will take it all down to evens if necessary rather than messing about and getting caught out.

As Derek says wisdom comes into it, a £200 dog race where just the fanatics are involved might produce a more accurate price than England vs Germany where huge numbers are making partisan judgements based on nothing but hope or loyalty.

Been thinking about this. On the flipside, theres no incentive for a team of people to analyse and work on a prediction model for greyhounds because their costs cant be met with the low liquidity.

However, a team of 10 data scientists making a prediction model for England Germany where they could maybe get £500k of bets on...

There are a lot of dog races

[andy28]

I think it was Peter that said its easier to make $1000 from 1000 races than $1000 from one race.

As a side note where is Shaun, I have not seen him on forums since Nov, hope all is well

[Kai]

Re market accuracy, forget maths let's just use logic. How do you work out the exact chances of something winning on a given day? You can't, And that's why the highest rated horse in the race won't be 1.01 even though it's the best one, so should win. The 'market' will be saying something like 'this horse should win half the time so it's priced at 2.0'. On that day it might be a 1.01 dead cert if all the others have been sick but on another day it's true price might be nearer 10.0 because it's just not in the mood for it. That's where the pro value punters make their money, comparing the prices on offer with their estimate of the outcome on that day.

Bookies don't generally need to get early prices very close because they'll only let you have a tenner on at those prices. Once the markets get into full swing and prices are being set by the hive mind, then they'll just do a basic accountancy job controlling their estimated sp liability by turning the taps on and off on certain horses with their prices. If they've had enough money on the fav they can slash its price, because they don't accept lay bets, that stops the money coming over the counter for it and income starts to come on other horses. Trackside bookies aren't much different, they're unlikely to see enough business to create nice balanced books so almost everything they do is offered a bit worse than the exchange and the apprentice just lays it off (backs it) on the machine. That's why I think prices often spike to round numbers, someone's just getting things hedged and doesn't want to risk being 'clever' drip feeding bets in. If they take a bet for 2 grand at evens the lad will take it all down to evens if necessary rather than messing about and getting caught out.

As Derek says wisdom comes into it, a £200 dog race where just the fanatics are involved might produce a more accurate price than England vs Germany where huge numbers are making partisan judgements based on nothing but hope or loyalty.

Been thinking about this. On the flipside, theres no incentive for a team of people to analyse and work on a prediction model for greyhounds because their costs cant be met with the low liquidity.

However, a team of 10 data scientists making a prediction model for England Germany where they could maybe get £500k of bets on...

Not sure market size reflects much on "accuracy", I've often found the bigger the market the slower it is to correct itself, as it needs a bit more confirmation

The amount of uncertainty around the current price should tell you more about how accurate it's likely to be, but it depends on market phase and some markets never fully mature

I suppose you can trade this uncertainty, or obsess over accuracy, but if late money is more bearish that should also tell you more about the psychology behind that market