Market Efficiency and Framing a Market

[James1st]

Hi 74.5.

There are 2 statements here. Firstly some people claim that Betfair markets are efficient. My comments related directly to the efficiency of the process by which a horses odds reach SP, viz a trading journey that is most certainly not efficient. Whether or not there is any desire to research methods that are more efficient isn't worth the debate since the Betfair process of trading to an SP price is all that currently exists and that is likely to continue as the preferred method by which SP is determined.

The second statement relates to the accuracy of SP and there is absolutely no correlation between the SP arrived at in Betfair and the efficiency of the process. My conclusion was that the process itself is inefficient because whether you choose a route through Aberdeen/Devon or you choose to go directly to Manchester, either party arriving at Manchester town hall demonstrates accuracy but not efficiency.

Moving on to the accuracy of SP's; it is obvious that taken globally, all SP's do in fact correlate very well to actual results. But individually SP's can be and are very wide of the mark and it all depends on how you formulate the subsets of the global view. For example, the subset of SP's that select 3yo Hcps only will also demonstrate that SP correlates very well with actual results. However, if you exclude from that subset all the horses where the jockeys are wearing orange shirts or orange caps, then the correlation between SP and actual results doesn't look so accurate.

Now, if individual SP's can show a great deal of deviation from the norm and subsets of the global view are also subject to error then one has to conclude that there are forces at work that manipulate global SP's to coincide with actual results. I am not talking of some secret society here , but I am pointing a finger at those that have most to lose by allowing average SP's to outstrip actual results.

The fact that global SP's can be demonstrated on a graph to equate to winning chances may well be an indesputable truth but it is certainly one that masks the large number of anomalies that exist below the surface.

[herbie]

If I was to back 5000 home favourites in the football, 100 times each randomly in the 24 hours before kick off and the same favourite 100 during the game (also randomly), would I end up losing or gaining anything?

If the markets are efficient I should be level minus commission. But hold on a minute my commission is 5%... some of you are on 25% and a few are even on 60%.

So is the market defining the true % chance of Newcastle winning at home over the long term or is it pricing the true price plus an average commission?

Going back to randomly backing pregame and inplay.

It seems that the two things that move a market are lack of money and lots of money... I could be wrong (as I am constantly told I am by my family and friends)..so the news or opinion creates this money situation ... please correct me if I am...

My point is that the pre-game price is pushed back and forth pre-game within a set range covering 0.?% to 10%+ sometimes without any news. So is the range x% working within the commission rate? Or is there long term value?

For the pregame price to be efficient doesn't it have to be right any given moment base on all available information. But what about Bully money that pushes the price back and forth and fear and greed etc etc

On Peters Blog yesterday he priced man city at 2.0 I spent the last hours before kick off scalping it. It sat around 1.97ish then off upto 2.0... a few grand taken 2.02 for the luck ones then it's backed into to 1.88. last few seconds I got matched between 1.88 and 1.91.. Which price is right?

In play is the same, a fav corner goes up the odds on fav can come in 3 to 5 ticks.. hows many active users know the true chance of that ball going in? Can excitement and desperation move this price? and which price is efficient?

This is happening thousands of times a week.

There are so many ways to price a football game.which one is correct? Also the efficient market should give me a price based on history and available news. If Man City are are 2.0 and they lost their next 4 games their price would alter... therefore the efficient market (if it exists) has to be organic and move based on the perception of the future.... at sometime even in it's purist sense will the price be wrong during it's transition?

I hope you understand.

herbie

[Euler]

From space the Earth looks round, but if you are a foot of Everest that is obviously not true.

[CaerMyrddin]

If the markets are efficient I should be level minus commission. But hold on a minute my commission is 5%... some of you are on 25% and a few are even on 60%.

The comission is paied on winnings, so it won't afect the way the market is priced.

[JayBee66]

...

Efficiency in the economic sense is the ability to discount current information into the price.

At any one time the price represents all that is known by the crowd and its wisdom.

If new information enters the market then that information will be digested by the crowd and may or may not move that price.

SP is regarded as the most efficient price because the race is about to begin and everyone that had an opinion has made it with their bets.

When the race starts the prices jump about as horses vie for position. As the winner crosses the line its price is 1.01 because the market has discounted that information (it won) into the price.

In precise terms the horse race betting market is referred to as "semi-strong form efficient" ie information takes a while to be discounted but discounted it is.

Efficiency in this case is not mechanical efficiency nor the efficiency of Betfair to match layers and backers on the order book. It is just the crowd's ability to discount information into the price.

Prices move around in the 10 mins afore the race start as the bookies, connections and insider traders lump their money on and the crowd takes stock.

By taking the SP and matching it against actual outcome you get the graph I posted above. The two lines almost match thus proving market efficiency. There is a favourite/longshot bias due to backers avoiding odds-on favourites and chasing losses with long priced no-hopers. Other than that, the market is semi-strong efficient by any economists standards.

Try using Google's Scholar search engine to search for papers on the subject. Many economists use the horse betting market to cut their teeth for financial market research as they are analagous to one another. Search for names like Ziemba and Hausch and follow the papers they cite in theirs.

It's a fascinating subject.

Now, the efficiency of the place market is something else entirely but I'm saving that for my MSc.

James Primero.

[74.5]

Hi 74.5.

There are 2 statements here. Firstly some people claim that Betfair markets are efficient. My comments related directly to the efficiency of the process by which a horses odds reach SP, viz a trading journey that is most certainly not efficient.

James
I don't think that the process is efficient.In fact, its is down right inefficient.But,given that information is coming into the betting environment all the time and the fact that there are various types of punters (backers,layers,traders,those with inside information,those without,those with large banks,those with small banks etc.),it isn't at all surprising that markets don't achieve efficient prices through an efficient process.In fact,I'd be shocked if the process was efficient.

Now, if individual SP's can show a great deal of deviation from the norm and subsets of the global view are also subject to error then one has to conclude that there are forces at work that manipulate global SP's to coincide with actual results. I am not talking of some secret society here , but I am pointing a finger at those that have most to lose by allowing average SP's to outstrip actual results.

My feeling on this subject is that the manipulation of global SPs would be difficult to achieve and would self-correct in any case.My research indicates that this does happen on the micro, but not the macro, scale.The 'manipulation' tends to be indirect rather than direct.Nonetheless,it is effective.I'd rather not say any more on this particular subject.
74.5

[herbie]

At any one time the price represents all that is known by the crowd and its wisdom

The point I was trying to get to, was that if someone is using money to push the the price one way or another by say 5 ticks over, say, one minute, then it's nothing to do with the crowds opinion....it's on person distorting the price. So by doing this constantly is the market constantly inefficient?

[JayBee66]

The point I was trying to get to, was that if someone is using money to push the the price one way or another by say 5 ticks over, say, one minute, then it's nothing to do with the crowds opinion....it's on person distorting the price. So by doing this constantly is the market constantly inefficient?

There have been many people trying to push the price of commodities by weight of money. They have all failed and bankrupted themselves.

You can buy up a range of prices on Betfair but if the market still thinks that the first price you matched is still the right price then the market will be back after you have your fun.

The statistics show that SP prices are efficient because they represent the true odds over time.

[JayBee66]

James
I don't think that the process is efficient.In fact, its is down right inefficient.But,given that information is coming into the betting environment all the time and the fact that there are various types of punters (backers,layers,traders,those with inside information,those without,those with large banks,those with small banks etc.),it isn't at all surprising that markets don't achieve efficient prices through an efficient process.In fact,I'd be shocked if the process was efficient.

I don't understand what you mean by "efficient process". I think there are many in this thread who do not understand the meaning of efficiency in economic terms.

This has nothing to do with mechanical, electrical efficiency, nor the magazine Health and Efficiency.

As I stated above, the market is semi-strong efficient. Information enters the market, it takes a little time to be understood by the crowd but understand it they will.

The race starts at a given time. Those that have their opinion on the race have to bet before the race starts. When the race starts the SP represents all that the pre-race market knew about the event. Statistics show that this information is efficient because the odds represent the true chance of winning.

What more is there to understand?

[Iron]

Statistics show that this information is efficient because the odds represent the true chance of winning.

What statistics are you referring to in particular?

What more is there to understand?

Quite a lot...

Let's say you have 1000 horses with average true odds of 5.0. 500 start at 7.0 and 500 start at 3.0. Overall, the market is extremely efficient on one level, but extremely inefficient on another.

And however you slice it and dice it, the fact that some people are able to make a living from trading shows that the market is full of inefficiencies.

Jeff

[Iron]

BTW, you guys might like this article on market efficiency: http://karmatics.com/docs/evolution-and ... rowds.html

Jeff

[JayBee66]

Statistics show that this information is efficient because the odds represent the true chance of winning.

What statistics are you referring to in particular?

What more is there to understand?

Quite a lot...

Let's say you have 1000 horses with average true odds of 5.0. 500 start at 7.0 and 500 start at 3.0. Overall, the market is extremely efficient on one level, but extremely inefficient on another.

And however you slice it and dice it, the fact that some people are able to make a living from trading shows that the market is full of inefficiencies.

Jeff

1) Any analysis of a large number if results will show that SP equates to real chance of winning.

2) You said, "Let's say you have 1000 horses with average true odds of 5.0. 500 start at 7.0 and 500 start at 3.0."

What exactly do you mean by "start"? SP? If the SP is 7.0 then the true odds are 7.0 not 5.0 True odds are SP. You cannot take averages of horses with different SP.

Yes, people are able to make money from localised variations in prices. Before the off a price will vary. Timing is everything. If the price is 5.0, 10 minutes before the off and you jump on it just afore it shortens to 3.0 then you have what the old-timers call value and what new-timers call a winning trade. The horse was deemed to be 5.0 by the market, new information came in and said that the actual price was 3.0

[Iron]

So are currency pairs, if you think about it, as otherwise the arbsters would have a field day!

Jeff

Sports are more efficient because the market has to be priced at 100%.

[Iron]

1) Any analysis of a large number if results will show that SP equates to real chance of winning.

Any analysis?

How many such studies are you aware of, and are you confident that they employed sound methodologies?

What exactly do you mean by "start"? SP?

No, BSP.

If the SP is 7.0 then the true odds are 7.0 not 5.0

But what if drifts frequently push BSPs into value backing territory (as a study by Betfair a few years ago suggested happens)? That was what I was driving at, ie that the drifts and the steams may cancel each other out, creating an impression that the market is extremely efficient (when on one level it definitely isn't!).

Yes, people are able to make money from localised variations in prices.

So would you agree that the Random Walk doesn't apply during those periods (as otherwise it wouldn't be possible to profit long-term in such markets)?

Jeff

[JayBee66]

So would you agree that the Random Walk doesn't apply during those periods (as otherwise it wouldn't be possible to profit long-term in such markets)?
Jeff

That is what semi-strong form efficient means.

You should read what I wrote in my earlier posts and not what you wanted to see.