Useful Statistics for Discretionary Trading?

B

BOC

I'm not designing a system, but as the thread title indicates, I am interested in useful statistics to get a better grasp of the instrument I'm trading, so I assume this is the correct forum. Right now I'm exploring some stats in Excel for the EUR/USD - the average high-to-low range for each hour, avg HL range for the early London session (3am-8am EST), avg HL range for the London/US session (8am-12pm EST), avg swing in pips and time on the 5-min chart (using Zig-Zag), avg volume during the aforementioned session times.

I'm new to stats and certainly no Excel expert, luckily these are pretty simple formulas. And I realize an avg is just an avg, still I think it might give one some insight into typical mkt behavior.

What other stats might a discretionary trader consider that might be useful?

Thank ye, thank ye.
 
BOC said:
I'm not designing a system, but as the thread title indicates, I am interested in useful statistics to get a better grasp of the instrument I'm trading, so I assume this is the correct forum. Right now I'm exploring some stats in Excel for the EUR/USD - the average high-to-low range for each hour, avg HL range for the early London session (3am-8am EST), avg HL range for the London/US session (8am-12pm EST), avg swing in pips and time on the 5-min chart (using Zig-Zag), avg volume during the aforementioned session times.

I'm new to stats and certainly no Excel expert, luckily these are pretty simple formulas. And I realize an avg is just an avg, still I think it might give one some insight into typical mkt behavior.

What other stats might a discretionary trader consider that might be useful?

Thank ye, thank ye.
More...
LBR tests patterns similar to the simple numbers you have. If you listen or read her work she will throw out nuggets from time to time. Try YouTube for starters
 
BOC said:
I realize an avg is just an avg, still I think it might give one some insight into typical mkt behavior.
More...

An average isn't typical, and it is very misleading. Try plotting the distribution of the metrics you are interested in, and for that be sure to group your data points in buckets of equal size (eg is you are looking at an hourly range, use buckets in 10 pips increments : 1-10 / 11-20 / 21-30 / etc) so that you can easily interpret the visual display of that distribution.

Get familiar with the notion of standard-deviation, in many cases you can use it to estimate the % of data points in-between avg +/- N * stdev.

Also, the information ratio (avg./stdev) gives you a quick idea of the relative shape of that distribution, and whether the avg is representative or not of anything ( the smaller the information ratio, the more flat is the distribution, and useless is the average)
 
Thanks Cswim, I'll check her out.

dom, yes, I've read that std dev is better, which I was considering. I'm somewhat familiar with the concept because of volume profile, the bell curve and such. That's a bit more complicated in Excel but google is my friend. Thanks for the input.
 
You forgot the most important item: number of observations
Are you going back 3 days, 3 weeks, 3 months, 3 years ?
Which or all ?
 
It's five minute data, about 30 months. I thought I'd do stats for the whole 30, then 6 months and 1 month. Something like that.
 
If you are going to do that, then compute the RATIO of the one and six month results as compared to the 30 month stats.
 
You must log in or register to reply here.
Back
Top