Is Volume worth a pinch?

[rinconpaul]

I hate it when you read something along these lines, "odds are purely a function of betting volume!"...Nothing could be further from the truth IMO.

Now you'd reckon a two outcome event would surely be priced according to the volume placed for either outcome, but sadly it has very little to do with prices. Let's take the under/over soccer goals market. In the referred pdf (https://www.wiwi.uni-muenster.de/uf/sit ... 3-11_2.pdf) it demonstrates that despite there being a 50:50 chance of over/under 2.5 goals scored, the money is piled onto the overs on average 80:20! It's a matter of sentiment, people like to cheer on high scoring outcomes, not boring non scoring events. Despite this unbalance prices generally reflect a 50:50 outcome. Vol analysis achieves no advantage!

The logic of volume affects price sounds good, but rarely is the case. Below is a dog race, betting for the Place. You can compare VOL$ (1/(individual volume/total volume)) to BACK$. There's no correlation whatsoever between BACK and VOL$ or between VOL$'s as a probability outcome. The two dogs that placed had the least amount of money on them. This isn't a one off, like I've scoured results to prove a point, it's rampant throughout all markets.

[Euler]

I'm not really sure what that publication was trying to research or why?

In the old days bookmakers lines used to follow their liability in the market but nowadays it's priced the other way around. People know the price and bet to value thereby eliminating any bias (in theory).

[rinconpaul]

I say forget volume and get to know RANK. Of the limited amount of info available from an exchange RANK is way more useful when paired with price? As for "people know the price", yes that's what that research paper proves, but it also proves that sentiment/bias, loyalty to a club etc can sway people to bet on an outcome rather than bet without prejudice, and NOT move the price. That sort of turns physics on it's head? The waters of exchanges are so muddied by traders antics, who are without sentiment and just following technical signals, and very little to do with fundamentals.

Price, and its implied probability, is not consistent either when you couple it to RANK. Question then: Does a 4th price ranked runner @ $2.50, have the same implied probability as a 1st price ranked runner @ $2.50? Traders probably couldn't care less, just whether it's in an up/down or hold mode and how many seconds to green/red up. Punters, on the other hand, should familiarise themselves with the concept if seeking VALUE!

[rinconpaul]

Another indicator that can be gleaned from the exchange is the % of new matched price changes as a % of total matched price changes per race.
More explanation: If you record prices at a fixed interval for a fixed amount of time pre race, up to start, and remove duplicates you might end up with something like this:
Runner 1......10 different matched prices
Runner 2......14 " " "
Runner 3......6 " " "
Runner 4......8 " " "

Runner 1 had 10/38= 26%, R2 = 37%, R3 = 16%, R4 = 21%

Let's say Runner 1's average matched price is $2.50 = 40% Implied Probability. If you collect enough data and then correlate these average prices with % of different matched prices, you start to realise that all
similar Prices are not created equal as regard implied probability! There's a sweet spot and each side a noticeable reduction. What it says is, that when new matched prices are created the price is steaming or drifting. Too much drift or steam indicates an underlying uncertainty about the runners prospects? Just the right amount, keeps it on track to perform as per it's Imp Prob Price.
In the graph I've plotted the Implied Probability of an average matched $2.50 runner in a 6 dog Place market (Y axis). The X axis is the % of different matched prices of the total matched prices of all 6 dogs. The sweet spot is between 17 & 24%. Anything more or less, and you are not getting Value.