Predicting volume one day ahead

[Gringotts Bank]

I'm starting this thread to ask if anyone has tried predicting volume one day ahead, and has had some success.

Either that, or perhaps you can suggest some variables to input/optimize.

Thanks.

 

[tech/a]

Yes
In SOME cases both volume and bar structure can be pretty accurately predicted.
Have done so for both entry and exit.
For arid net sake a pivot point reversal bar will generally herald a down bar on above average volume.

When back from holidays and charts in hand I don't mind doing a bit of EOD live type analysis around this.

How do you intend to use it?

 

[Gringotts Bank]

tech, I'd like to be able to predict when tomorrow's volume has a high probability of being +50% on today's volume. Intention would be to trade intra-day, soon after open on a select bunch of speccy stocks.

 

[Mistagear]

GB,

Will be interesting to see if you can derive an edge by predicting high volume likely follow, without there being a skill-set which in itself gives a greater understanding of the market, thus gaining greater advantage anyhow.
High volume can act as a confirmation factor more than a cause imho, so am thinking likely more useful to be able to identify the cause ahead of the resulting volume.

It's my contention that all volume is supply, therefore low volume can sometimes be a more powerful precursor to price advance.
Additionally, very often high volume is needed prior to a direction change, in the form of soaking action to exhaust supply. In this instance you might be better knowing, after the fact.
Perhaps I'm off in the wrong direction to your intention with this, happy to gain new understanding.

Cheers, M

 

[tech/a]

MG
Don't disagree with some of your obsessions and haven't tried this before.
Will be interesting back on deck mid July.

 

[Tanaka]

I anticipate that knowing whether there will be high volume will save time when day trading. Waiting around for hours for a setup wastes time and makes day trading quite a boring profession. I mostly day trade the ES for this reason, I get the same setups with the SPI but I’ll be in and out within 15 minutes on the ES where as I’ll be in and out in 2+ hours on the SPI. But does high volume equal high volatility?

Anyway, if you could predetermine high volume and scan for it could produce more setups. Great idea GB! I look forward to Tech getting back and sharing his ideas, selfish bastard travelling the world :

 

[tech/a]

I anticipate that knowing whether there will be high volume will save time when day trading. Waiting around for hours for a setup wastes time and makes day trading quite a boring profession. I mostly day trade the ES for this reason, I get the same setups with the SPI but I’ll be in and out within 15 minutes on the ES where as I’ll be in and out in 2+ hours on the SPI. But does high volume equal high volatility?

Anyway, if you could predetermine high volume and scan for it could produce more setups. Great idea GB! I look forward to Tech getting back and sharing his ideas, selfish bastard travelling the world :

Sitting in a park outside of ORLEAN France.
Perfect.

It's simple but complex in terms of finding setups.
As MG has pointed out much is supply but we can get an indication of when supply has dried up.
I've done a fair bit of work on extremely low volume bars and I'm sure you'll find the answer to the riddle there.

Rather than looking for high volume bars look for EXTREME low volume bars and up to 5 bars before to see if it's supply or demand drying up.

I think for the ES you may need 15 to 60 min bars to see it clearly --- could be wrong I've only traded very little in currencies.

But here is where I think the search should start---away from the obvious.
An increase in volume will come EVERYTIME you get an extremely low volume bar--- the trick is ---- is the volume lack of supply or demand.

Answer that and you have your setup.
Now dinner.

 

[Gringotts Bank]

But does high volume equal high volatility?

Tan, I suspect it will translate into high volatility when applied to the right selection of stocks. I have a few scans I can use for that aspect of it.

The tricky bit will be backtesting, because I will have to "buy" a stock's volume. Probably need to look into custom backtesting and user defined metrics. Anyone done this sort of thing before?

Mista, I read high volume as "market interest", and whether it's buying or selling interest, a lot of people are watching. I like that in a stock.

 

[Wysiwyg]

Either that, or perhaps you can suggest some variables to input/optimize.
Thanks.

Pretty basic but for an increasing volume event with a greater than 50% possibility, what about looking back at volume for the number of occurrences there are over e.g. 5 years

1) ref(v, -1) > ref(v, -2) & ref(v, -2) > ref(v, -3);
2) ref(v, -1) < ref(v, -2) & ref(v , -2) > ref(v, -3);

1) the number of times that volume had increased for two days equals 100
2) the number of times that volume increased for 1 day and decreased the next day equals = 500

then the probability of volume decreasing the next day after an increase is more likely by 5:1.

Adding an increase above MA or an increase in multiples or an increase in % are variables.

 

[tech/a]

While I've not completed detailed analysis PEN would be one I'm on I posted the trade in #1510 post in the PEN thread it had a volume setup that suggested the move it had in the end ---a few days after my post.

Worth watching going forward
The stop is the low of the most recient double bottom.

 

[Gringotts Bank]

Thanks for the input wys, but can you explain a bit further please. I'm not following. How does one make condition 1 = 100 and condition 2 = 500?

 

[samanne1]

Thanks for the input wys, but can you explain a bit further please. I'm not following. How does one make condition 1 = 100 and condition 2 = 500?

Just need to look at the conditions graphically (tho drawing horizontal cause I cant be bothered to insert piccy)

example 1 is like this

day -3 xxx
day -2 xxxxxx
day -1 xxxxxxxxxx

whereas example 2 is like this

day -3 xxx
day -2 xxxxxxxx
day -1 xxxx

(just twist your head 90 degree to see it the right way)

the references to 100 and 500 only relate to the number of time that pattern was found. Think it was only meant to be an example of detecting increasing volume.

Cheers

 

[Gringotts Bank]

right thanks sam.

This look right? I guess it's not unexpected to have results that are <.5 for most stocks. What else can I try - anyone?

a = ref(v, -1) > ref(v, -2) & ref(v, -2) > ref(v, -3);
b = ref(v, -1) < ref(v, -2) & ref(v , -2) > ref(v, -3);
d = cum(a);
e = cum(b);
f = filter = d/(d+e);

addcolumn(f,"probability of 3rd day V increase");

 

[Gringotts Bank]

What conditions should I look at today to predict tomorrow's volume?

What would make most sense? Breakout?

I only want to know if the volume is likely to be higher or lower than today's.

 

[tech/a]

Is your question rhetorical?
I don't think you can predict EVERY day
But quite a few.
Put up a list of codes and we can see how good I think I am.
Better than 50 % would e significant I feel.
60% + would be above chance I would suspect.

Hang on I gear ge kettle boiling got a LOT of tea to make

 

[Gringotts Bank]

Just having an earl grey myself...

Let's say CEO, NWE, LEG, CGM.

 

[tech/a]

Just having an earl grey myself...

Let's say CEO, NWE, LEG, CGM.

CEO not one Id trade.
But Monday up or neutral 43c most likely immediate destination.
NWE wouldn't trade that either.
Best opportunity was at the bottom of the recent consolidation.
Small rise then reversal could be in the same day.Short term blimp.
LEG nor this.
4c area max move see more consolidation here. But not the worst.
CGM you have my interest.
Only problem is the lateness of finding this one.Great break from long consolidation and was easy to see.
One that we all want to find!!
I see this ranging a little and would like to see a few more days of price action before I took a trade.
I feel the low will be tested as will the high.At those points decisions will become clearer. So up to neutral.

 

[rnr]

I only want to know if the volume is likely to be higher or lower than today's.

Tech,

I think this is the question GB wants answered.

 

[Gringotts Bank]

Tahnks rnr.

I see CEO Monday having volume < Friday.

The others I can't tell.

 

[rnr]

I see CEO Monday having volume < Friday.

Okay, so we both share the view that the volume will be down from Friday but how will you trade CEO given that belief (if it helps then assume you could also short the stock)?