Dennis's quote "Assure you never miss entry into major trend"

How do you know when a trend has started? Ended? This forum is for discussions about trend indicators and signals.
Post Reply
zoopy12
Full Member
Full Member
Posts: 15
Joined: Mon Apr 05, 2010 7:50 pm

Dennis's quote "Assure you never miss entry into major trend"

Post by zoopy12 »

Hello Bloxards, Recently I saw a trend in the gold futures and it evolved in a way that did not allow me to participate according to my personal and highly defined objective trend system. The next day, Gold did offer me an entry, but that entry and exit was not nearly as profitable as the prior move that my system missed. This caused me to start thinking about something that I have heard many times in reference to the Turtles and Richard Dennis. One of his key axioms was something like this: Whatever method you use, the most critical thing is that if there is a major trend, your approach should assure that you get in that trend. For some reason I cannot see clearly the logic of HIS statement even though it might be true. My thinking has always been that my system identifies a subset of all possible trending expression shapes of market movement, and that it should on balance make money, when only MY set-ups and entries are taken. However, when I consider that the ATR of each day (market opportunity) is made up of ALL kinds of price expression (including mine as a part of them) and I add the Richard Dennis statement to that, I am left with the nagging suspicion that I might be extracting only a fraction of the day's ATR potential, by not participating in those other trends, that a "Dennis type" system might capture. I do certainly believe that this question/dilemma CAN be answered by pure logic alone and does not need a computer backtest. Could someone please offer some insight in the simplest terms. Perhaps, we can do it by saying: Here is Richard Dennis's statement (stated above) Is it true (useful) or not true(not useful), and just as importantly, why is it true or not true? I welcome any insights. Thanks, the Zoop
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Re: Dennis's quote "Assure you never miss entry into major trend"

Post by Chuck B »

I've always taken that statement in a very simple form. He's saying that if there is going to be a large move in a market (many, many multiples of ATR from the entry), he wants to ensure he is in that movement. Again, he's looking at daily or weekly data "trends" that for such moves oftentimes can last many months to a year or more. It sounds like you're more focused on short term stuff, but within the context of any system, you likely want to ensure you catch the big winners that are presented in some way; full well knowing that everything is 100% data dependent and rules you have that filter entries based on massive historical testing, etc, can and will not function in the future as you might prefer.
zoopy12
Full Member
Full Member
Posts: 15
Joined: Mon Apr 05, 2010 7:50 pm

Re: Dennis's quote "Assure you never miss entry into major trend"

Post by zoopy12 »

Hello Chuck, Thank you for very clear insights here. It certainly helped. I guess the thing for me to do is to mess around with the charts and see the difference between my more narrow approach to trend following versus a "blind" price breakout method and consider the pros and cons of each. Who knows I might end up with a combined method by using my (pattern based approach of swing highs and lows with my entry filter) + a fail safe price breakout entry and stop, and simply execute whichever one occurs first. I might need a third hand and a second brain but automation is getting better all the time I hear. Thanks again, Zoop
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Re: Dennis's quote "Assure you never miss entry into major trend"

Post by Chuck B »

The exit strategy and position sizing strategy are vastly more important than the entry strategy. If a market is going to trend for the next 14 months, just about any trendfollowing entry strategy will get a signal very early on. The key then of course is what initial and ongoing exit rules are applied and how much risk, as a percentage of total equity, is assumed in the position at every step along the way.

The meaning of the word "trend" goes back to a recent discussion on the board where I pointed out that it means something different to many people. If a market moves from say 1300 to 2400 over 14 months, in hindsight most would agree it was in an uptrend given it appreciated ~85%. However, it's very possible that "trendfollowers" could lose substantial sums over the 14 months all depending on their exit rules.

Let's say said market achieves that 85% appreciation but experiences numerous 12ATR pullbacks along the way, maybe 50+ day lows, etc, then resumes upward...perhaps it does such with a spike downward in today's famous illiquid fashion over just a few days time. Then it immediately resumes the upward bias (while in fact many "trendfollowing systems" are now short the market). So it comes down to how sensitive the exit strategy is, how much leverage is carried, etc, and how sensitive a reentry strategy is to the price data.

Unfortunately, there's no free lunch here. An ultra-long trend trader who uses weakly or monthly price data doesn't bat an eye while the typical "trend" system would be chopped to pieces through an 85% price move that features multiple 10-15ATR off the peak retracements.

Some in-progress examples: Palladium had an almost 12ATR decline, reached a 71-day low, 19% off the March peak, and almost every traditional trend system would have exited at some point if not attempted a short entry. Is it's uptrend over? No idea. Nobody knows. It's back near the Mar high on a continuation chart.

TU (2yr T-note): Uptrend since Nov low would have needed a 6ATR trailing exit to maintain the position. Same with the 5yr. 10yr just scrapes by with a 5ATR trailing exit, same with long bond. Dec20 Eurodollar STIR would have needed a 6ATR trail to hang in this uptrend so far also.

A trend of beauty has been the Euribor since the Oct low (Dec2020). A 4ATR trail has captured this beautiful price action without issue. In the meantime, the 2yr Schatz has needed a 5ATR to capture its vastly more ugly/less rewarding price action. However, stepping out to 5yr Bobl shows a beautiful low volatility about the mean uptrend move that a 4ATR has easily contained.

Note that the "typical" trend system if there is such a thing generally has issues staying in a position with adverse moves greater than ~4-5ATR off the extreme of the move. Many are gone by 3ATR off peak. By 6ATR most all have long ago puked out the trade. For that matter few traders can emotionally accept a 6ATR exit which allows a position to potentially give back a couple of months or more of gains while still hanging in there.

So...it's all in the exit and how much heat is used in the portfolio. Entries are almost meaningless over long term trends (at least that is a useful bias to assume) since if you get onboard that market headed to 2400 at 1300 or 1340, 1380, etc, it's somewhat academic.
Roger Rines
Roundtable Knight
Roundtable Knight
Posts: 2038
Joined: Wed Oct 06, 2004 10:52 am
Location: San Marcos, CA

Re: Dennis's quote "Assure you never miss entry into major trend"

Post by Roger Rines »

zoopy12 wrote: Tue Jun 25, 2019 2:48 pm [SNIP]...Recently I saw a trend in the gold futures and it evolved in a way that did not allow me to participate according to my personal and highly defined objective trend system. The next day, Gold did offer me an entry, but that entry and exit was not nearly as profitable as the prior move that my system missed.
[SNIP]...I welcome any insights. Thanks, the Zoop
Consider this idea:
"Some types of systems perform better when they are enabled to be active for a fixed period of time. This means that once the system creates a signal, the signal stays active until a different signal is created, the signal is filled, or the signal expiration period expires.

How much time a signal should stay active can be estimated with testing.

Keeping a system's signal persistent for a set period, or a new signal is established, enables a system that is missing some favorable moves to be enabled. Adding a persistent signal to an existing system provides a mechanism the system’s short term signal-life missed. Often the reason it gets missed is the conditions for a repeat of the earlier signal didn't have the setup validation values a new signal required to catch the earlier move. "

Roger...
zoopy12
Full Member
Full Member
Posts: 15
Joined: Mon Apr 05, 2010 7:50 pm

Re: Dennis's quote "Assure you never miss entry into major trend"

Post by zoopy12 »

Wow, 2 very powerful posts, Chuck and Roger. Let me step in with a Mea Culpa right off the bat. I just went through and tallied all of my paper trades and I don't fudge my paper trades, for the last 6 months in three different symbols YM, GC, CL, my system has done horribly. Chuck you were right on with saying it's a short term system, it is. Each of the symbols had at least 300 trades. They are all in significant deficit. None of them has seen daylight. There is no pressing of winning trades under any condition, and the golden rule that I have heard says, you cannot press(add lots on positive movement) your way out of a losing system. So that's my big self revelation. My method is pattern based S/R buys High of day sells low of day breakouts and it has a fairly stingy initial stop, but I suspect that the losses are more from the lost opportunity of big trends that materialized but my entry could not get me in or got me in very late. I think the phrase is "efficiency" of some sort. I mean in the last 20 days there has been this amount of total price movement in points, what did my system capture of those "available" points. I am thinking/almost sure my system does horribly on this entry part. Hence the reason for exploring the turtle method for my entry. On the exit side of the equation, I probably have some "inefficiency" also, and that would probably stem from getting stopped out on a pattern violation and having no Turtle type reentry. So I end up watching and waiting during X points of positive movement in my original direction,while my pattern develops X points higher. In the past, I equated re-entry with Revenge trading, so I purposely made my rules for it extremely narrow, My pattern at a new high/low of day. Whereas I think now,if I get stopped out, I must have a way to get back in on a new high/low of day. Also what scared me, beyond the revenge notion was the question of how many signals am I going to end up taking today as the account is not huge. That is what lead away from the Turtle type methods many months ago. Fortunately we now have liquid Equity Micros so this opens up a whole new range of options. And yes, Roger I was thinking along the lines of what you mentioned, although I did not think in terms of a time window of signal validity, I just thought If I am going to take Non normal re-entry signal it has to meet criteria X which was usually some minimum level of proximity to the last high or low of day, the idea being that the whole momentum may have shifted and my previous stop out was with good reason. And Chuck thanks for the rigorous exit observations. Fortunately, I have absolutely NO qualms with watching winning trades drawdown, pretty severely, just to stay in the trend. I have seen the white paper from a group that talked about using I think 10R trailing stops on stocks long only, and showing great success with that approach in position trading, I think Crittenden was the firm ?? On this note I would recommend an eccentric fellow named Guy Fleury who proposed a shocking approach to long term trend foliowing equities and he used Berkshire Hathaway as an example, where the position becomes geometric and Totally self financing after the initial outlay of I think 200 shares at 10$. At the end the snowball is in the hundreds of billions, if I remember. Quite remarkable. In the end, I think I am going to have to admit that Turtle Style trend following is probably my best approach, whether that can work in shorter time frames is an open question but I don't see why not at this point. Thanks for great insights, Zoopy
Post Reply