ATR Stretch Candidates 12-15-10
Here are the results from the Scan:
CLICK TO WATCH LONG IDEAS VIDEO:
CLICK TO WATCH SHORT IDEAS VIDEO:
Please watch this video where I explain why
CLICK TO WATCH VIDEO
CLICK TO WATCH PART 2 of VIDEO
12 Responses to ATR Stretch Candidates 12-15-10
Interesting findings regarding ATR. I only use it to determine position sizing (not for long/short trade ideas), but the inconsistency in data between providers is quite alarming. While how I apply ATR has not caused me many problems, I may consider redesigning my position sizing metrics to use historical volatility instead. Nice work.
Reply
That’s basically what I’m working on, replacing ATR with HV.
If you only trade off of one data feed, there probably won’t be a problem because you will always be consistent. But the data difference is as you said alarming. The same problem is experienced with an exponential moving average, but the degree of error is much smaller as to be more trivial (but there nonetheless)
Reply
Congratulations on the props from StockTwits, Danny.
Well earned.
Fame and TMZ are just around the corner!
Reply
thanks a lot Anton. It’s nice to get recognized
Reply
Hi. I wounder if you got an email adress?
I would like to ask u some question regarding your “thoughs on Bpop& Volatility”, which u mentioned in on of your screen cast
sure, it’s spydercrusher “AT” gmail “dot” com. Just replace the at and dot with the appropriate characters.
Reply
Best blog on the internets! Thanks for the hard work. Keep the SF videos coming.
Reply
thanks a lot Dr.
Reply
Merry Christmas Danny!
Reply
I just got through the first video that speaks of the inferiority of the ATR calculation. You had stated that this was most likely due to there being more data available to one charting package versus the other. I noticed that you had both TC and SF set to a lookback of 14, Would it be that their calculations are fubared instead of it being data availability? That is indeed odd…
now on to the second video 
have to wonder if this strategy would fare better when taken with the trend. ie. only take shorts when you have stacked descending MAs ( 200 > 50 > 20) and ascending MAs on the long side (20 > 50 > 200). or perhaps some other trend qualifier…
Reply
the trend qualification is built in the algorithm bc it states the price must be 2*atr above or below a selected MA. In this case I used 21.
That’s also why I tried to heavily disclaim this strategy bc it picks tops and bottoms, it was more for intellectual satisfaction, thought there were insights gleaned (esp with respect to volatility grouping).
Strategies that picks tops and bottoms are typically very difficult (both in implementation and on the gut). I prefer strength and trend following bec there is a more logical way to program those things.
Reply