I had an interesting conversation (short but interesting) with James from Traderunner.pro about stats.

I have / had a stat that says that if price opens in value then there is an 83% chance of it testing yesterday’s vpoc. He asked whether I had considered whether that stat is more influenced by the distance from vpoc when it opens rather than fact it opens in value.

This has been something I had noticed before, that distance / size of various could be an influencing factor in the probabilities. For example, the 83% also includes all the times it opens just a few tick from the vpoc which certainly skews the results. Many of my stats have certain filters that take in account this sort of behaviour.

I have now applied that filter to majority of my stats that certainly alters the results. For example the open in value and retest of the previous days vpoc drops to 60% , if the open is more than 20 ticks away from the vpoc.

One of the problems with stats is that how do we use them practically to gain an edge. We could have a stat for fading a certain move but how do we find entry and how do we define our risk.

One of the techniques I have been trialling is an idea, I borrowed from @trader1906Â from his and Tom Dante‘s webinar on market statistics. My review of that webinar is here.

After I find a high probability (70%+) stat, I then look at all the times this stat has completed (thanks to RT Investor this is pretty simple) and looked at maximum distance it went in the opposing (adverse) direction. I then worked out the Mode distance and the 1 Standard Deviation of this maximum adverse distance.

What this allows me to do, is that when a stat becomes valid but open, I can look at the chart for valid levels within the max adverse distance to take a possible trade with a target of completing the stat. I then have a way of defining my risk and knowing that if price trades beyond this standard deviation that whilst the stat still have a chance of completing, that the probability of this has dropped to where it is now not a high probability setup.

I have been using this technique on a couple of stats and these have now turned into my most frequent stat trades and my most profitable stat trades.

This is the chart of my filter and the MAE profile, for the days where it only breaks one side, how far did it go back inside the IBR when the extreme is outside the IBR.

### Like this:

Like Loading...

*Related*

Hi Adey,

Found your post interesting, especially the last part where you have the profile of the MAE of your setups, I’ve been trying to do that with IRT but struggling a bit, would you mind sharing the txt file so we can see how it is built ?

Regards,

David

Hi David,

Thanks for commenting. I will post the txt file for one of my charts so you can see how it done.

It will likely be Thursday as I am off to London tomorrow, and have a bit on at the moment.

Adey

Hi David,

My apologies for the delay in replying. I have posted an answer to your question here.

http://takingonetradeatatime.com/2016/07/02/return-stats-follow/

Regards

Adey

No worries, I took me some time aswell to answer ! I’ve figured out how to do it myself, it took me a few hours though. I also saw the method Hugh used, a bit different but the results are the same ! Careful though in the Breakout failure part 2, if I recall correctly there is a SESST which is not set to the correct “Ignore the first X minutes” of the session. It’s set to 240 but it should be set to 300 I guess.

Regards,

David (@pookehz)

Hi David

Thanks for the comments. I did notice that in Hugh chart. There is many ways to skin a cat. đź™‚

Regards

Adey