The short answer is I am not trading on now but I have built models I expect to trade on at some point where volume is important.
I think the paper I posted in the other thread is a good example of how I think about it. The author's basic strategy is that large weekly price moves tend to reverse. He has a story about how volume might be related to the probability a reveral happens and how strong it is. Two stories actually but only one seems true. Then he tests it in historical data for a large number of stocks by making portfolios of winner and loser stocks with different levels of volume increases/decreases in the prior week.
I have repeated this exercise in a different timeframe and stock universe with good results. Like your experience it is not a guarantee and might be hard to see trade by trade, but comparing a portfolio formed on the basic signal to a portfolio formed on the basic signal plus a volume rule it looks like a real edge.
If your trade setups are rule based and you have the data, you could just compare you P&L distribuions over subsets of trades taken with different volume conditions, maybe it matters and maybe it doesn't. But I like the idea of starting with a plausible story of why volume might be related to future returns for a particular type of stock and then testing it.
In the paper, the story is that reversals are stronger if the initial move was not based on information and trading volume increases when information is in the market.
I can think of two other volume stories that are out there. One is that if you can start with a set of price moves that are not information based (probably not as easy as it sounds), larger volume moves are liquidity based and will have stronger reversals. The other one is a bit more complicated - people make decisions differently if their position is in a gain or a loss, and looking at past trading volume at different prices can help identify stocks with large unrealized gains or losses.
<B>
--------------------------------------------------------------------------------
Quote from Thunderdog:
But that's the point, isn't it? Sometimes price continues on higher volume and sometimes it reverses. Sometimes price continues on relatively low volume and sometimes it reverses. Have you been able to distinguish in advance which is which? Personally, I have not been able to do so with any practicable reliability. So the question I had to ask myself was, what is the incremental value of this additional variable?
As an aside, let us exclude relatively large spikes in volume from consideration, since the writing is on the price wall as well when such outsized volume spikes occur, with outsized price bar ranges. So I don't think there is incremental informational value in such occurences as it relates to volume. As for notable increases in volume not accompanied by outsized price bar ranges, only the subsequent price action will tell you what's what, since I have found that it can go either way in my own experience with no practicable a priori bias that I, personally, can distinguish. So where, then, is the value in volume? Perhaps your perception is more nuanced in these matters.
--------------------------------------------------------------------------------</B>
