Market Efficiency

[EyePeaSea]

I'm not a newbie in terms of time on this forum, but I think this is a newbie question so...

In this forum, I've seen people mention market efficiency. Is that meant as an assessment of how accurately the market reflects a runners likely performance? I understand that both exchange and bookie odds will reflect where the money is going, but also the odds can reflect peoples perception of how likely the runner is to win. So, is an efficient market one that more accurately matches the likely outcome than an 'inefficient' market? If so, is the feeling that BF is getting more efficient over time or less?

Thanks

Ian

[LeTiss]

I used to be very confident in the market efficiency - a guide as to how accurate a selections chances of winning

However, over recent times, that market efficiency has been distorted. There is so much manipulation that goes on these days, not just Betfair, but also Joe Bloggs with his automated bots

[EyePeaSea]

Thanks. So I was broadly right about the meaning, but you feel that if anything, the efficiency/accuracy may be getting worse.

My strategy involves backing X number of the most favourite horses. If an outsider wins (say, a 10/1), then I go red. Interestingly, based on my 'what-if' statistics, the market seems to be slightly more "efficient" about 2 hours from the start than closer to the start time (but it's a close thing).

I think I might look at the data a bit more. See what the odds are (and their position compared to the other runners) of the eventual winner are at 2 hour, 1 hour, 30 minute and 1 minute snapshots before start time.

Thanks again.

Ian

[LeTiss]

Thanks. So I was broadly right about the meaning, but you feel that if anything, the efficiency/accuracy may be getting worse.

I have no evidence to back that up, Ian

It's just a general feeling i have when trading the markets these days - I'm not 100% certain the market is as efficient as before, as there's so much manipulation now

That doesn't mean, they cannot be traded - there's still plenty of opportunities and money to be made. But, even BA's excellent automation is another example of efficiency being impacted upon. I've even placed momentary large stakes myself, 2 or 3 ticks away - then watched how it immediately triggers bots based on WOM. That's not an efficent market, that's pure manipulation, even on my part.

[Euler]

I used to be very confident in the market efficiency - a guide as to how accurate a selections chances of winning

However, over recent times, that market efficiency has been distorted. There is so much manipulation that goes on these days, not just Betfair, but also Joe Bloggs with his automated bots

I'd say that was about right, the markets are more efficient on a broad scale because they have to be as the aggregate sum of all possibilities has to add up to 100%. But the variability in the markets now is high, so they can not be fully efficient.

I collected some data on both recently so was going to blog about it at some point.

[kerberus]

You can also play around with "Last Traded Price" try picking a 50/1 geegee and tickling it with 10p bets!

[EyePeaSea]

Thanks to both of you.

Ian

[spreadbetting]

I'm not a newbie in terms of time on this forum, but I think this is a newbie question so...

In this forum, I've seen people mention market efficiency. Is that meant as an assessment of how accurately the market reflects a runners likely performance? I understand that both exchange and bookie odds will reflect where the money is going, but also the odds can reflect peoples perception of how likely the runner is to win.

These days a lot of late money is now betting on how a runner will perform in running rather than simply how likely it is to win. I think you'd need to drill down the data rather than making assumptions as so many things can tip bare stats into profit but might not neccessarily be exploitable.

[firlandsfarm]

Hi all, I'm fortunate enough to have a database of UK and Irish racing going back many many years. I have BSP from outset. I thought it may help this Thread if I published a table of BSP (rounded to a Betfair Tick odd) showing the number of Runners at each price, the number of Winners, the Expected Winners based on the price, the Expected Win% based on the price, the actual Win% and the required Win% after allowing for 5% commission (the yellow column). So to take you along the BSP 3.00 row for example ...

There have been 2070 runners in all races with a BSP of 3.00 (that will actually be between 2.99 and 3.025 because of rounding).
Of these 675 won their race.
Compared with an expected number of winners of 690 [ races/BSP ]
The expected Win% is 33% [ 1/BSP ]
Compared to the actual win% of 32.61% [ winners/races ]
Which can be compared to the required Win% of 34.48% after allowing for the 5% commission charge [ 1/((BSP-1)*0.95+1) ]

So I think it's reasonable to say the market is as efficient as could be expected. I just knocked this together in a rush ... I will republish it year by year 2008 - 2016 when I get a little time so that we can see if there has been a noticable trend.

I am not intending that these numbers are used to come up with a system! Any advantage/disadvantage of actual wins over expected wins seems pretty random in fact I put in an Advantage column but it was so random I took it out again.

[EyePeaSea]

Thanks for the post and putting the effort in. Lots of data