I'm not sure if this question should be here or in technical analysis. Does anyone know how a volume histogram study is programmed? I know what it is suppose to be showing, but I want to know how it is programmed. I use Ensign software and I want to write an ESPL script for a custom study, but I can't get Ensign to give me a clue as to how their code is written to create the study they have. Since it displays off of historical data for any give time frame, I doubt it is truly displaying accurate volume at individual price levels and since it doesn't actually display the volume level value, (only a vertical line representation), there is no way to verify whether it actually is able to determine the volume totals at each individual price level. I have been able to write scripts that capture the Bid volume versus the Ask volume and I want to try and apply that data to a volume histogram study. Any clues, comment or advise would be appreciated.
We have something like what you may be looking for in our alerts software: http://www.trade-ideas.com/Help.html#HRV
When you say "volume histogram" are you talking about a volume at price (also called a price profile) study? If so, it can be computed from the raw trade T&S data (which might be time consuming to do for a heavily traded stock). I've also seen it approximated without diving into T&S by using a distribution of each bar's volume across the price range of the given bar and then summing across to produce the volume at price study values.
Below is part of an email I sent to Ensign about their volume histogram study and the last paragraph is Ensign's response. "As far as the existing volume histogram, I have my doubts as to whether it actually is showing correct volume at specific price levels. Mainly because it is displaying it on historical data where as the bars it is displaying against are only giving OHLC and total volume for a give time period. How is this study able to total up volume at individual price levels? For instance, when displayed on a 3 minute chart and a given individual bar may have a 3 point range (for the ES), which would be 12 individual price levels, how is the current standard study able to give an accurate volume for each individual price level." "ANSWER: That's not how the study is calculated. For each screen level (pixel position) that the Volume histogram plots at....the study determines how many bars are intersected at that price level. If a Bar can be intersected at that price, then it is included in the study for that price. The same bar is used several times (for its complete High / Low range)." Now I ask you, does anybody think that this method would give you a decent approximate volume at each price level? Do you suppose this is how all charting programs create a volume histogram study?
While a kind of absurd way to do it (by using intersecting pixels), it sounds like it would be similar to the equal distribution of a bar's volume method I mentioned earlier. Since the volume distribution method seems to produce similar results to using the actual T&S (at least on 1 or 5 minute charts anyway - could significantly skew as time period increases) one could probably assume the "pixel" method would work similarly. The pixel method wouldn't compute any volume - it would only give a relative result.