New Members: Be sure to confirm your email address by clicking on the link that was sent to your email inbox. You will not be able to post messages until you click that link.

2nd filter: OBV

Dear Mark D,

To identify bullish price breakouts, I use a Slow Stochastic crossover. Then, monitor the behavior of the Slow Stochastic for 3 days.

I want to add a second, volume-related filter to my initial scan. Then, monitor the behavior of whatever I used for 3 days.

Although there are other volume-related indicators, I would like to start by using the OBV. Should I use the OBV, the OBV Signal(n), or both? What would be the settings?

Thank you!

Doctor T


  • markdmarkd ✭✭✭
    I don't use it, so I can't help you with settings. The original intent of OBV was to get "early warning" of changes in the longer term trend, usually by divergence - in other words, say in a down trend, price would make a new low on the current down leg compared to the last down leg, but OBV wouldn't follow. That suggested the selling was weak, so if buyers came in, the stock could rise. Vice versa at tops. But we're talking about trends of several months. It probably could work more short term, too, but probably not as well.

    I couldn't find an explanation of OBV Signal, but appears to be just an MA of OBV, so, like any signal line, OBV crossing above the falling MA would be bullish, crossing below the rising MA would be bearish. What's the right length for the MA for Signal? Who knows. Different ones will work better in different markets. You just have to try it out on stocks you are interested in. Stockcharts seems to like 30. I tried a few lengths, 21 seems to be reasonable. But, if the stock goes into a range, you will get whipsawed, and if it moves sharply up or down, signals will be late.
  • markdmarkd ✭✭✭
    If you were thinking of using an OBV signal crossover together with a Stochastic crossover in the same scan, you would probably get few hits. They are unlikely to happen simultaneously.

    But, you could try scanning for one crossover, say, stochastics, and then filter for OBV > OBV Signal. Or, vice versa, OBV crossover its Signal, with a filter %K > %D.

    Another alternative, scan for the Stochastic crossover, with OBV below Signal, put the result in a list, next days scan that list for OBV cross above Signal. And vice versa.

    Just things to try. Don't know if they will work.
  • Dear Mark D,

    Thank you for your replies!

    I am going to do some backtesting.

    Doctor T
Sign In or Register to comment.