OmniVest User Forums
General Discussion Topics
Using Stops With RTMs
7 replies, 1355 viewings
|
OmniVest User Forums
General Discussion Topics
Using Stops With RTMs
|
Last Activity 4/15/2019 7:43 PM 7 replies, 1355 viewings |
|
|
|
| Printer friendly version |
Mark Holstius Elite    Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL  User Profile |
This review started with the 9,378 trades generated over the past 15 years by my FTM portfolio. I obtained the historical stock data to identify the lowest low for every symbol for the duration of each trade and used that to determine if exiting at a stop during the trade was beneficial or not. I used Stops ranging from -1% to -20%, and the results are tabulated below; The data in Black cells refers to the original FTM stats. The P/L% sum of all original individual trades was 14,165% (this is NOT a measure of CAR). The columns contain statistics for the various Stops ranging from -1% to -20% in rows. Since it’s a popular number, I’ve highlighted the row for a Stop set at -10%. 1) Of the 9,378 trades, 264 were exited early because a low during the trade triggered the -10% Stop Loss. 2) 55 of those exits were beneficial and resulted in an increase of +153% to the 15 yr P/L% sum of trades. 3) 209 had a negative impact. If those 209 trades had been allowed to continue they would’ve had a P/L% better than -10%. Those “missed opportunities” would’ve added an additional 1,641% to the sum of trades vs the -10% Stops. 4) When that missed potential of 1641% is considered, using a -10% Stop resulted in a decrease in the P/L% sum over 15 years of -1,488% (12,677% vs 14,165% in the original sum), along with a small decrease in the HR. I ran the same analysis with my larger dataset of 25,118 unique RTM trades over the past 15 years with similar results; 835 Stop Exits at a stop of -10%: 242 beneficial and 593 with a negative impact. These statistics are for this particular set of trades. Even though it’s a large dataset, it’s unique to those trades (a curve fit). While using stops may be beneficial with longer duration strategies, these tests appear to support the idea that Stop Losses may not be helpful with short term RTM trades. In every case, any positive impact from getting out before a loss was overshadowed by the negative impact of missing trades that were more profitable than the stop. I’m currently expanding my previous work on Limit Orders and hope to have more information on that in a few days. As always, I welcome comments and suggestions. Mark [Edited by Mark Holstius on 8/31/2017 7:35 PM]  Attached file : 01 FTM Stops.png (211KB - 776 downloads) Attached file : 02 FTM And ARM4 Stops.png (227KB - 691 downloads) | ||
| Jim Dean Elite  Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
As you know, I've done a pretty massive amount of development work on different types of stops over the last 15 years or so. I think they are, in general, more important to "get right" and "apply intelligently" than getting the "perfect entry" is. JMHO, but I'm not alone. Without providing a loooong discussion about whys and wherefores, let me just state that I have *never* found that price-percentage based stops are optimal ... there always are better approaches (simple and complex) My "simple" recommendation is that ATR-multiple based stops are nearly always better than percentage based stops, except maybe for long term (multiple months) trading. So, I'd suggest you rework your evaluation using atr multiples for stops, rather than percentages. Obvious questions is "what multiple". Common-sense answer is that the multiple should be in "safe" proportion to the anticipated profit (ie consider Reward vs Risk). If profitable trades average, say, 2 ATRs, then the stoploss certainly shouldn't be more than 2 ATR. Also, the logic of the stop needs to be designed with the entry logic and typical progress of the trade in mind. Nirvana's approach to RTM's, by my observation, usually attempts to enter while the pullback is still occuring, rather than after it reverses. That means a lot of "faith" is needed when entering the trade, that the pullback is not actually a reversal. So, ye of "little" faith, if you want to "demand" that the pullback ends immediately after entry, and protect yourself from true reversals, use a fairly tight stop at first, then maybe loosen it after the reversal. That is, consider the low of the entry bar, maybe the LLV(3), as a "pivot point" that if not validated by an immediate reversal, should be the threshold for immediate exit. OTOH, ye of "much" faith (in N's strategy), you'll presume that an overwhelming majority of the entries will result in profitable resumption of the trend after a couple of bars, even if the wait time gets a bit iffy. In that case, do the reverse ... use a fairly wide ATR multiple for the first couple of bars but then seriously tighten it after the resumption occurs. Final statement ... as you've repeatedly and correctly mentioned, Mark, there is an unavoidable positive bias to the Excel testing you're doing since presumably there is *some* degree of curve-fitting in the strategy out of the box. So, use the statistics with care regarding the "little vs much faith" alternatives mentioned above. Hope this gives you some things to consider, maybe a bit more grist for the mill. | ||
| Mark Holstius Elite Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL User Profile |
Interesting idea, having the stop tighter the first day or 2 with RTMs to weed out the "failures". Hmmmm.... I totally agree that a stop based solely on Price % isn't the best - but the limitations for me of Excel & the amount of historical data needed to calculate ATR necessitated its use. Hopefully, N will give us better options...??? Mark [Edited by Mark Holstius on 9/1/2017 11:17 AM] | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
simplest TR formula = max(High,C1) - min(Low,C1) ... and of course you know how to do averaging :~) [Edited by Jim Dean on 9/1/2017 11:18 AM] | ||
| Steve Mayo Legend Posts: 414 Joined: 10/11/2012 Location: Austin, TX User Profile |
The other thing that's not considered, however, is that once stopped out, those funds then could be put into new trades which potentially could have better prospects. So, another way of looking at stops (instead of as a protective measure), is to think of using them as a money management approach that, hopefully, moves us quicker out of low-potential trades back to our system's normal-potential trades. Steve | ||
| Mark Holstius Elite Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL User Profile |
Thanks, Mark | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
I believe that well designed stop-logic is just as non-arbitrary as well designed entry logic. Extreme example - I have a "superstop" that has 3500 (efficient) lines of code. Not at all arbitrary! Most non-arbitrary simple but useful stops only require 50-250 lines or so, btw. And I agree 100% with Steve's money-management comments! That's yet another reason why well-designed stops are important. Stops should not only end (or scale out of) a trade when losses mount, but also if profits stagnate when price flattens, and also when statistically extraordinary positive gains have been experienced. [Edited by Jim Dean on 9/2/2017 5:11 AM] | ||
| Buffalo Bill Member Posts: 28 Joined: 1/2/2015 Location: Stafford, VA User Profile |
as always an excellent analysis. nice to feel comfortable knowing that using the fixed loss stop is a bad idea vs wondering each time we get a "large" loss trade |
|
|
|
|
Administrator
Forum Moderator |
Registered User
Unregistered User |
 E-Mail this thread to a friend(un)/Freeze thread | |
Toggle e-mail notification |
