Storing Traded Volume Data

[hecojef928]

It would be nice if there were separate columns for back and lay data, enabling me to track traded volume on each side. This breakdown could provide better insights into market dynamics. I've encountered a similar limitation in Guardian where back and lay volume can't be saved as stored values, which would be quite beneficial. Including features like a Historic Relative Volume Condition would be a useful. Given its simplicity, it's surprising that this is missing. Its incorporation would offer a clearer picture.

While I'm able to store the last traded price and the volume traded at a specific price, there's no distinction between back and lay volume. I've heard it's roughly around 50%, is this correct? The lack of differentiation hampers my analysis. It seems someone else has raised a similar question /viewtopic.php?t=25318. However, I do believe it's a valid suggestion. Having a clear understanding of traded volumes and POC would certainly help.

For now, I will look into Dallas suggestion based on an average.

[Dallas]

As discussed in that topic you linked to its always 50%, if the traded volume at a price is £4234.50, that has to be made up of £2117.25 back money and £2117.25 lay money.

But if you are trying to differentiate between if an amount matched was lay money taking back money or vice versa then as Shaun that harder and not very accurate, especially when you factor in XM'ing and XMXM'ing in football markets

[hecojef928]

As discussed in that topic you linked to its always 50%, if the traded volume at a price is £4234.50, that has to be made up of £2117.25 back money and £2117.25 lay money.

But if you are trying to differentiate between if an amount matched was lay money taking back money or vice versa then as Shaun that harder and not very accurate, especially when you factor in XM'ing and XMXM'ing in football markets

While it's helpful to understand that matched volume is evenly split, I'm of the opinion that a volume profile would provide me with a deeper understanding of the dynamics, especially in Horse Racing. Being able to examine the volume distribution across different price levels would offer me more insights? To avoid getting stuck in analysis paralysis, I'm eager to hear from yourself and others on this.

[fingers21]

On a similar note, is it possible to store the traded volume of a selection above or below a certain price and if so how?

[hecojef928]

On a similar note, is it possible to store the traded volume of a selection above or below a certain price and if so how?

Are you asking about adding a separate column to the ladder, or are you referring to the stored values in Guardian? If you're talking about Guardian, wouldn't the "Volume of a Selection" or "Volume Traded at a price for a selection" area within the Stored Values be what you're referring to?

[fingers21]

I am asking about stored values. Basically I am trying to figure out how to add the traded volume of a selection over a certain price and save it as a stored value to be used in an automation rule.

So, as an example, a selection is currently trading at 4.0 and the traded amount at 4.2 is £1000 at 5.0 is £750 at 7.0 is £600 and at 9.0 is £420. Therefore the total traded above 4.0 would be £1000 + £750 + £600 + £420 = £2770.

In the above example I used only 4 prices but in reality the number of prices with traded volume could be more or less.

[jimibt]

As discussed in that topic you linked to its always 50%, if the traded volume at a price is £4234.50, that has to be made up of £2117.25 back money and £2117.25 lay money.

But if you are trying to differentiate between if an amount matched was lay money taking back money or vice versa then as Shaun that harder and not very accurate, especially when you factor in XM'ing and XMXM'ing in football markets

While it's helpful to understand that matched volume is evenly split, I'm of the opinion that a volume profile would provide me with a deeper understanding of the dynamics, especially in Horse Racing. Being able to examine the volume distribution across different price levels would offer me more insights? To avoid getting stuck in analysis paralysis, I'm eager to hear from yourself and others on this.

i've created a c# api implementation and am weirdly just tackling this very topic as we speak. I likewise feel there is merit in understanding the price/match distribution model and am similarly focussed on the pre race horses in this endeavour. will share what i find as i'm certain that sharing the mechanics of HOW it works won't diminish any edge.

[jamesedwards]

I am asking about stored values. Basically I am trying to figure out how to add the traded volume of a selection over a certain price and save it as a stored value to be used in an automation rule.

So, as an example, a selection is currently trading at 4.0 and the traded amount at 4.2 is £1000 at 5.0 is £750 at 7.0 is £600 and at 9.0 is £420. Therefore the total traded above 4.0 would be £1000 + £750 + £600 + £420 = £2770.

In the above example I used only 4 prices but in reality the number of prices with traded volume could be more or less.

I don't think there is an option to check the total traded volume above a certain price. So you need to add up the traded volume of all prices at 4.0 and above. I would use this kind of method to achieve this.

test.baf

[fingers21]

I think that may do what I need. I will have to play around with it a little bit but it looks promising.

Many thanks for your help!

[decomez6]

As discussed in that topic you linked to its always 50%, if the traded volume at a price is £4234.50, that has to be made up of £2117.25 back money and £2117.25 lay money.

But if you are trying to differentiate between if an amount matched was lay money taking back money or vice versa then as Shaun that harder and not very accurate, especially when you factor in XM'ing and XMXM'ing in football markets

While it's helpful to understand that matched volume is evenly split, I'm of the opinion that a volume profile would provide me with a deeper understanding of the dynamics, especially in Horse Racing. Being able to examine the volume distribution across different price levels would offer me more insights? To avoid getting stuck in analysis paralysis, I'm eager to hear from yourself and others on this.

i've created a c# api implementation and am weirdly just tackling this very topic as we speak. I likewise feel there is merit in understanding the price/match distribution model and am similarly focussed on the pre race horses in this endeavour. will share what i find as i'm certain that sharing the mechanics of HOW it works won't diminish any edge.

The effect of liquidity on volatility is greatly influenced by crossmatching . This tend to happen across the whole market. so unless you factor the latter , match distribution is just another lagging indicator.
most traders will concentrate on selections with high 'Old volume' which in return create value on long priced ones with less volume.
BUT you must know when to predict the correlation especially when there is too much heat on the front end of the book.

[hecojef928]

i've created a c# api implementation and am weirdly just tackling this very topic as we speak. I likewise feel there is merit in understanding the price/match distribution model and am similarly focussed on the pre race horses in this endeavour. will share what i find as i'm certain that sharing the mechanics of HOW it works won't diminish any edge.

Thanks! Best of luck!

[jimibt]

While it's helpful to understand that matched volume is evenly split, I'm of the opinion that a volume profile would provide me with a deeper understanding of the dynamics, especially in Horse Racing. Being able to examine the volume distribution across different price levels would offer me more insights? To avoid getting stuck in analysis paralysis, I'm eager to hear from yourself and others on this.

i've created a c# api implementation and am weirdly just tackling this very topic as we speak. I likewise feel there is merit in understanding the price/match distribution model and am similarly focussed on the pre race horses in this endeavour. will share what i find as i'm certain that sharing the mechanics of HOW it works won't diminish any edge.

The effect of liquidity on volatility is greatly influenced by crossmatching . This tend to happen across the whole market. so unless you factor the latter , match distribution is just another lagging indicator.
most traders will concentrate on selections with high 'Old volume' which in return create value on long priced ones with less volume.
BUT you must know when to predict the correlation especially when there is too much heat on the front end of the book.

the heat factor is prescient!! would love to garner more of your perspective obviously, but am aware that we all chew the stick in different places and at varying times.. i'm currently (in my new interest in this stuff -pre-racing) looking at:

1. slope() across a number of defined data points
2. standard deviation on the same(ish) subset of data
3. rate of change (roc) of price (but using a balanced book% to compare)'

Initially, i was going to use machine learning and use the obvious features of time, price, slope, roc etc, etc, but have gone for a more trad approach (simpler) and am purely looking at the 3 states above as linear comparison. thus far, not much to report as i've only begun collecting the data but am confident that i can make sense of it based on those 3 simple data points mentioned above.

in short, regards activity on previous prices, i'm thinking that i can either look at those prices as being dead, OR being prices of interest.. will keep you posted.

[sionascaig]

the heat factor is prescient!! would love to garner more of your perspective obviously, but am aware that we all chew the stick in different places and at varying times.. i'm currently (in my new interest in this stuff -pre-racing) looking at:

1. slope() across a number of defined data points
2. standard deviation on the same(ish) subset of data
3. rate of change (roc) of price (but using a balanced book% to compare)'

in short, regards activity on previous prices, i'm thinking that i can either look at those prices as being dead, OR being prices of interest.. will keep you posted.

That's interesting & something I've been looking at too (in general terms)...

I discounted using a standard deviation as don't think its of much use in a live situation, take for example a price that only moves in one direction (it has no std deviation) but found the rate of change to be of value. And rather than standard deviation it may be worth looking at the traded range over different time intervals & measure volatility that way.

But to cut a long story short I ended up jumping to the end game and asking the question if the market does X where do I want my bets to be. So about 10 months into it with a bot and results are very promising.

Barley scratched the surface though and as well on ongoing refinement of the X bot need to pick up the Y, Z, A , B, C... scenarios too )

[jimibt]

But to cut a long story short I ended up jumping to the end game and asking the question if the market does X where do I want my bets to be. So about 10 months into it with a bot and results are very promising.

Barley scratched the surface though and as well on ongoing refinement of the X bot need to pick up the Y, Z, A , B, C... scenarios too )

i think that's exactly how i see it, i.e. WHERE do you want to offer, rather than WHAT

I discounted using a standard deviation as don't think its of much use in a live situation, take for example a price that only moves in one direction (it has no std deviation)

which interestingly puts you on the offer for the price coming to you!!

... but found the rate of change to be of value.

that i feel is the confluence piece in all of this. marry up some of the above and determine where the push is coming from. As i'm sure you're aware, the prices can all be boxed up into a *container* which has a max volume of 100(%). Understanding this can allow you to see which parts are moving in unison (think pistons) and which are purely utility to the market.

[jimibt]

Barley scratched the surface though and as well on ongoing refinement of the X bot need to pick up the Y, Z, A , B, C... scenarios too )

i know you quoted Dallas on scalping before in a previous thread, but it's worth surfacing this again as it still makes sense in 2024:

I dont do much scalping myself but when i do my general guide is.

*Don't go race to race trying to scalp everyone the sameway instead be selective and look for the race types and times that best suit this method.
*Avoid all Maidens, Novices 2yo races etc
*Avoid races with short priced fav's
*Look for a competitive h'cap or the better quality races which have the front few in the betting priced similarly (ie 4.0 - 7.0)
*Target runners trading in a tight range and ignore anything thats been trending in a direction (unless directional scalping) or has a betfair chart that looks like a heart rate monitor - or you may end up needing one!.
*Wait till the market is turning over at a decent rate somewhere around a £200 per second average increase, the quicker you can be in and out or in and scratch the better.
*Pay close attention to the previous race going in-play and especially to when it finishes - at these times there will be a influx of money as people start to take positions and this can change everything - even runners that have been rock solid all day could move several ticks at this time and never come back so best just holding back around these periods till things settle again.
* Use the Risk Meter on the Bet Angel over view screen this will give a clear indication to the overall activity - the lower the better when scalping.

and when your not doing that keep reading this forum, the BA blog and watch more video's.