Good (Tuesday) morning all.
Well: it would appear that I’ve found something to keep my ‘idle hands’ busy during the day!!! Statistics!!! (Although: I’m sure you’ve all heard the anecdote: ‘There are lies, damn lies, and statistics’)!!!
Well: I’m not sure if this information means anything or not BUT, using ADX as a guide, I thought I’d try to see whether or not there is any merit to the statement (question???) that I made in my post earlier i.e. do some instruments ‘trend’ better than others???
Well (simply put):
I coded an ‘indicator’ that counted the number of days that ADX was below +DI AND -DI OR ADXR was below 20. This ‘indicator’ also counted the total number of days (bars) on the daily chart. By then simply dividing the total number of days (bars) on the daily chart by the number of days that ADX was below +DI AND -DI OR ADXR was less than 20: I arrived at what I called the ‘ADX Range Factor’. The theory here was: the HIGHER the ‘ADX Range Factor’ the LOWER the amount of days that the instrument traded ‘sideways’ i.e. had little or no directional movement which, as we know, is what ADX indicates to us.
Now before I discuss the results and before you look at the attached document:
In certain cases the results MAY be ‘skewed’ for the simple reason that some instruments have not always been available for trading at Delta so the numbers of days (bars) in the daily chart is LESS than a ‘full’ chart (which is 999 bars). Now whether or not this, in the long run, will make any difference to these results we shall have to wait and see. In other words: it’s glaringly obvious from the results that a lesser number of days used gives higher ‘ADX Range Factor’ results. Having said that: it’s the RON/TRY/RUB pairs that have fewer bars than the rest BUT they have the SAME number of bars as EACH OTHER so THEY are being compared to EACH OTHER ‘like for like’. I HAVE previously mentioned that those particular pairs appear to ‘trend much better’ and my results appear to confirm this FOR NOW i.e. given a ‘full’ chart I don’t know if this will still hold true.
I used ADX(14). I am happy, however, that no matter what period you use, because ADX / ADXR is calculated INDEPENDANT of PRICE, you would get the same results.
Now for the results:
(As I said: I don’t know if they mean anything because I’m not a statistician. Having said that, according to ADX anyway, certain instruments really do appear to behave differently over the same given period of time).
For all the instruments where there was a ‘full’ chart of data i.e. 999 days (bars) in the chart:
AUD/CHF, EUR/NZD, and (would you BELIEVE) AUD/NZD are ‘top of the pops’. The ‘ADX Range Factor’ for each of these pairs was calculated to be over 4 (the HIGHER the ‘ADX Range Factor’ the LESSER the number of days that the instrument had little or no directional movement).
Next on the list were USD/SEK, EUR/ZAR, GBP/ZAR, EUR/SEK, NZD/CAD, and EUR/NOK. The ‘ADX Range Factor’ for these pairs was calculated to be over 3. Oddly enough: when I was ‘playing around’ with the ‘Commodity Selection Index’: these paris were ALMOST ALWAYS ‘top of the pops’ with the CSI so I guess that this ‘test’ (at very least) confirms the CSI.
Then we get to the ‘other’ instruments:
The ‘ADX Range Factor’ for most of the instruments (forex pairs) that had ‘full’ daily charts i.e. 999 bars / days in the chart was calculated to be between 2 and 3. To be honest: I was quite surprised that the ???/JPY pairs fell into this category. CHF/JPY (it would appear) spend MUCH of it’s time trading with little or no directional movement. Oddly enough: this is the one pair where the ‘look of the chart’ has always ‘put me off’ i.e. far too many swings in opposite directions on alternating days as well as LONG shadows. I dont’ know why but this pair has always ‘spooked me’ and maybe this is why.
Next we have Gold and Silver:
Unfortuanately they have not been offered for trading at Delta for too long a period so whether or not that fact that they both have HIGH ‘ADX Range Factors’ (above 3) means that they in fact have the same directional movement as the forex pairs that are ‘top of the ADX Range Factor pops’ I cannot say.
Then we have NZD/CHF and AUD/JPY. Both have VERY high ‘ADX Range Factors’ BUT very few daily bars on the chart. I SUSPECT that by just looking at thse two pairs: you cannot really compare them with the others. NZD/CHF has a very high ‘ADX Range Factor’ but, as I said, the ‘sample data’ was lacking. Again: whether or not this makes a difference I do not know. I suspect that it DOES make a difference.
Then we come to my ‘beloved’ ‘exotics’:
Again: very few daily bars when compared to the other instruments (forex pairs). Each of them do, however, have the SAME number of daily bars so it’s safe to say that when comparing THEM to EACH OTHER: EUR/RON spends more time that the others trading with good directional movement.
Last: we have OIL!!!
Again: the number of daily bars available for the test was low but it has the highest ‘ADX Range Factor’ of all. Can this be compared with the rest of the information??? Would you be comparing ‘like for like’ or ‘apples with apples and oranges with oranges’??? I don’t have a definitive answer yet. I suspect that you cannot compare it with the rest of the information. Having said that though, on the other hand: the ‘ADX Range Factor’ DOES INDEED take into account that there are fewer daily bars in the sample data so who knows. I don’t!!!
Then I moved on the the indices (I have not bothered to include them in the document). They ALL have a ‘ADX Range Factor’ above 3. Can they be compared with the forex pairs that have a high ‘ADX Range Factor’??? They DO all have FULL charts (same as the forex pairs at the top of the list) so I’d say yes.
I then took a look at a few individual stocks. This (I think) is where is CAN get REAL interesting (these details are not in the document YET either as I have not completed any results). Interestingly enough: MOST stocks that I tested had the SAME ‘above 3’ ‘ADX Range Factor’ as the major indices (which, of course, does make sense). They also all had ‘full’ charts so I’m happy that I’m comparing ‘like for like’. There ARE, however, certain stocks, that have EXTREMELY high ‘ADX Range Factors’ i.e. some tested had ‘ADX Range Factors’ above 5!!! Now THIS could become interesting ESPECIALLY if one then looked at certain SECTORS!!! I’ll ‘keep going’ with this. You NEVER know what the end result could indicate!!!
As always: ALL comments are MOST welcome!!!
Regards,
Dale. (forexbrokersonline.net).
Edit:
I decided to attach two additional documents.
In ‘ADXRFRoundedSort’:
I rounded the ADX RF but still kept instruments seperate by the number of bars available for the sample.
In ‘ADXRFRoundedRFSort’:
I sorted the same rounded work sheet but ignored the fact that there was a different number of bars available for the sample i.e. this one is purely sorted on the ADX Range Factor.
Take your pick!!!
I SUPPOSE that ONE interpretation of these results could be that the instruments with the HIGHEST ADX Range Factors will generate greater profits when traded with the DMS???
By the way: I know it would APPEAR that I’ve now gone ‘dilly’ about ADX. I probably have BUT think about this: EVERY SINGLE SYSTEM in ‘the book’ REALLY is dependant on, or references, the current ‘state’ of ADX and ADXR not so???
As I know I’ve also said before: MORE DIRECTIONAL MOVEMENT EQUALS MORE PROFIT so it’s now very logical (to me anyway). If the markets did not move we’d not make any money??? Right??? And the MORE they move DIRECTIONALLY the MORE OFTEN we can use a TREND FOLLOWING SYSTEM and AS WE ALL KNOW (to coin that ‘inane’ phrase again): ‘The Trend Is Your Friend’ (I cannot stand it. It’s so ‘cliched’ but hey: am I willing to argue with it??? Nope)!!!
As I ALSO know (and have stated this week): when I’ve just been able to break even or lost money I’ve not bothered about the ADX ‘state’. When I’ve ‘chosen’ my trades based on the ADX ‘state’ I’ve always made good profit. There MUST be SOMETHING in THAT statement!!!
ADXRangeFactorCalculation.pdf (10.2 KB)
ADXRFRoundedSort.pdf (10.1 KB)
ADXRFRoundedRFSort.pdf (10.1 KB)