Probabilities of New High or New Low Across the Day

Another “sick day” another study refreshed. My pain is, hopefully, your gain.

In this study I examine whether there are any outliers in when each of the senior index futures tend to make new highs and new lows. It’s a pretty simple but very powerful study. It covers the ES (S&P 500), the NQ (NASDAQ 100), the YM (Dow Jones 30) and the TF (Russell 2000).

Methodology: I used 5-minute RTH bars from each instrument for the period October 20, 2011 back to October 19, 2010. This sample yielded 260 sessions, with 80 data points. As I ran through the bars, I simply counted each time a new session high or new session low was made during that bar.

Note that this is not necessarily the high or low of day, just each time a new high or low was made. In other words, one session can have multiple new highs and new lows. Lastly, I tossed out the first bar of each session because those are always new highs and new lows, so there is no point in counting them. Times on the x-axis are Pacific.

So here we go…

Before you peek below, what would you guess? Would you intuitively imagine that new highs and lows are randomly distributed throughout the day? Or would you figure there are patterns?

Lets look. Here are the probabilities of each 5 minute bar will be a new high:

Probability - New High

… and here are the probabilities each 5-minute bar will be a new low:

Probability - New Low

Some Observations

  • The 7:05AM bar is the proverbial big deal, especially for the ES, just like we saw in the NYSE TICK correlation study.
  • There is a cluster of new highs that occur just around to the 9:00AM bar. This is counter-intuitive at first, but when you think about the fact that European markets close at 8:30AM Pacific time, this makes more sense. Often, I’ve observed algos pause to let the European markets close to and then complete their programs afterward.
  • This is the biggest one, and one that I think you can integrate directly into your trading. There is a pretty strong time bias to new highs and new lows throughout the day. Again, as I noted in the tick correlation study, there is a tendency for new highs and lows to occur around the tops and bottoms of the hour.

Those are a few bullets to get you started. What else do you see? Trade ’em well.


Leave a Reply

Your email address will not be published. Required fields are marked *