Statistics and the IB Range

Trading the Bund (FGBL) I use the first 2 hours from the open at 0700 GMT for the initial balance range (IBR). As part of my trading, I use statistics to help me determine a possible intraday direction for my trades.

FGBL 2hr IBR Breakout

Previously over 1465 day the IBR has been broken one side or both 98% of the time. This is a handy stat to know but covers every IBR whether it is 10 ticks or 100 ticks.

After reading trader1906 tweets, I thought it worth doing a study on small and large IBR and what has happened in the past.

FGBL IBR <= 30 ticks

The above chart shows that in the past 2419 days there has been 1439 days where the IBR has been equal or less than 30 ticks and out of them 1439, 48% have broken both sides. So a coin flip whether it breaks one side or both sides.

I was interested in looking for a possible tipping point of when a one side break is favoured, I made a chart that looked at when IBR where greater than a set length.

great
FGBL >= 30 

The above chart shows IBR that are 30 ticks or greater, which now has a stat of 70% that only one side will be broken.

So if we have range of 30 ticks then we have 2 conflicting stats, one that says it 50% chance saying that both sides will be broken and the other saying there is 30% chance.

The problem is that all the less than stats includes the very small IBR and the greater than will include the very large ranges. Both of which will skew the results one way or the other.

The next step was to produce some stats for when the IBR was a certain size. The problem with having a specific size is getting enough sample size for these stats to mean anything.

I decided that I wanted approximately a sample size of 500 instances.

FGBL IBR >= 26 and <= 34

The above chart shows that a IBR of 26 to 34 ticks has broken both sides 38% in the past.

FGBL IBR => 35 and <= 50

The above chart shows that is the IBR is 35 to 50 ticks then it has broken both sides 31% of the time in the past 452 instances.

I looked at the small ranges and when the tipping point favoured both sides to be broken. It remained around the coin flip territory until we got under a range of 20 ticks and then it was a problem of find enough instances.

The range of 15-20 was still coin flip territorry but a range of 6 to 15 ticks produced a 62% that both sides would be borken.

What I learnt from this, is that the IBR is for me a coin flip whether it will break both side until the IBR is 30 ticks or wider and then only then should I be looking for one side of the IBR to hold over the a broken side.

I am no student of maths, and I could be approaching this all wrong.

If you have any comments then please free to tweet me at adeyf69 or leave a comment below.

Leave a Reply

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