Home > General Topics > Journals > Trend Trading to the MAX

Trend Trading to the MAX

  1. After discussing the topic on another thread, I decided I should start a journal.

    The trading used in this journal (experiment) is done to try and determine how much better your odds are of making a profitable trade when you trade "with the trend". I would like to quantify these odds as best as possible.

    I will use a real money account, and use about 20K in buying power to start. My stops will be set at 3% and my profit taking will also be at 3%. Thus, I can see how much the trend improves my chances.

    In theory, picking random stocks infinite number of times will yield 50-50 results as far as winners and losers. My current estimate is that trading with strong trends will yield between 60-65% winners. I will not discuss how I define a strong trend for a couple reasons. First being that this journal is not being used to critic my trading style, only to document it and discuss the importance/profitablity of trading with the trend. Second, a strong trend can be defined in many ways. You can look at the trades I make and hopefully agree that (in some form or another) a relatively strong trend exists.

    I will not have a time frame, the trades will remain open until either +3% or -3% is hit. I will trade with $5,000 blocks. If my expectations are correct, I should gross anywhere from $1,500 to $3,500 after 100 trades. It is harder to determine how long it will take to reach 100 trades. It could be anywhere from 4 months to a year.

    I will make an attempt to update daily, even if there is no activity.



    RT
     
  2. Just wanted to clarify what you are saying. $5,000 blocks each trade and 100 trades would mean about how much in starting capital? IN other words it is just $5,000 in one position at a time or you are going to use, for example, $20,000.

    I am interested because of the $1,500 to $3,000 gross and on what amount or buying power in order to folllow along.


     
  3. Op Coach


    I plan on making trades of $5,000. But I am only allocating $20,000 BP to this. So at most (in the begining) I would only have open 4 trades at a time.
     
  4. <i>The trading used in this journal (experiment) is done to try and determine how much better your odds are of making a profitable trade when you trade "with the trend". I would like to quantify these odds as best as possible.</i>






    your odds are no better. this has been tested rigorously.


    surfer
     
  5. Just curious what you would be using as your entry criteria? I know that you didnt necessarily want to discuss your trading strategy, but in this particular instance, only allowing 3% wiggle room either way could dramatically sway the overall results. Specifically, using stops based off of %'s, as opposed to technically defined stops, could hurt the performance in the long run. This is also where position size comes in.

    Just my .02

    jm
     

  6. Are you serious??

    If my odds are no better than why do people even pay attention to trends?
     

  7. Well, I use % so that the trend can be appropriately measured. If 3% turns out to be too tight, I may adjust in on a trade-to-trade basis. Either way, stop loss = target profit.


    EDIT: Will post first entries soon
     
  8. First trade already can and went.

    Long 350 STXS @ $14.469 on 1/20
    Sold 350 STXS @ $15.07 on 1/23

    Gross: + $210



    Current open position(s):

    Long 150 NRPH @ $33.55 on 1/23
     


  9. yes.

    trends only exist in the past.

    if you flip a coin 10 times and its comes up heads 10 times are you in a heads trend?

    good luck,

    surf
     
  10. You can have a system that makes money 30% or less of the time and still beats systems with 80-90% success rates because riding trends is what creates the profits, not the probability of the indicator to be right. So the best thing might be to define where in the timeframes these trends occur, and trade those timeframes. They usually last months, so no point looking for them in minutes or hours.

    Random entries can work, but youll find the implementation of that set up is difficult and your better off throwing stuff at the wall and seeing what sticks, generally something much more basic. To get random entries you need to scale in and out of trades, etc and it gets tedious. The best system Ive heard of with random entries only made 20% a year on a backtest. Tom Basso apparently used one in trading.
     
  11. Terrible analogy. The outcome of each coin toss is (for the most part) independent of previous results. People very often enter the market based on what has already happened- so it *is* very often dependent upon previous outcomes.
     
  12. Long 195 ESCL @ $25.60
     
  13. agreed. worst analogy i have heard in a while



    by your statement, surfer, it makes me believe that you fall into the category of traders who attempt to time the tops and bottoms.
     
  14. If you believe that there is absolutely no relationship between current prices and previous prices, as your coin toss analogy suggests, then there is also no basis for trading on statistical or mathematical criteria either, as your trading guru, VN, seems to do. Further, your much loved gann wheels (and/or squares) would have flat tires by now. (Which they probably do, anyway. Sorry, but I am not a fan of mysticism.) Finally, charts would then have absolutely no information value whatever aside from the current price. May I assume that you use no charting package, then? I would hate to think that you might be a hypocrite.

    As for your constant haranguing about the worthlessness of trends, have you not yet figured out that one man's trend may well be another man's countertrend? Therefore your glib conclusions hold about as much water as a spaghetti strainer.

    As for your comment that trends only exist in the past, that must have been quite an epiphany for you. In fact, they also exist in the present moment, as we all do. Nothing and no one exists in the future, and trends are no different.
     
  15. Further, I doubt that market complexity can be explained as simply as referring to mere joint probabilities associated with coin flips.
     
  16. i certainly do not wish to clutter this journal with the statistical/real life basis of my statement--UNLESS ofcourse runtrade wishes more------ with that said---- let me know..

    best,

    surfer


    ps. ALL the tests i have seen indicate a negative correlation, in all time frames, between past and future movement in the stock indexes. but please, don't allow that to stop you from trend trading.
     

  17. sorry that the coin flip analogy was over your head. perhaps, looking at this way will help you--- look at the coin flip as the second you enter the trade--- not as a market analogy per se.

    surfer
     
  18. Over my head? I studied statistics in university at both the undergraduate level and as part of the MBA program which I completed in '84. Did you? While I am certainly no statistician, I think I have the coin thing down pretty well.
     
  19. All those tests would have to assume a uniform working definition of a trend. In the real world, I doubt that we all assume that a trend starts at the exact same moment, even if we happen to be looking at the same data in the same time frame. Further, we would not all agree on what constitutes a sufficiently profitable trend, because we use different sizes of protective stops (reward-to-risk criteria). And that is even assuming that you are using price as the only variable, which is hardly always the case. You are trying to draw glib conclusions based on glib definitions.
     
  20. I interpreted your coin-flip theory to mean that the markets are totally random- meaning each tick is completely random- its result (whether it is an uptick or a downtick) is not at all influenced by previous ticks. If that is what your metaphor was- then I think you're are wrong. If you meant something else, I apologize.
     

  21. no need to apology, finn. differing interpretations of the data is what makes the market work. may i suggest "practical speculation" by niederhoffer and kenner for further insight of the ideas i am talking about.

    best,
    surf
     
  22. You could try to explain it here.
     

  23. not without permission from thread starter. sorry.


    surfer:)
     
  24. Neiderhoffer should write another book, how to write options and blow up twice.
     

  25. perhaps the next one will be "how to be ranked number one in the world after blowing up". don't think anyone else can lay claim to this success in the annuals of the market.
     

  26. I have to agree strongly with this - it is pretty tough to look at say a weekly or monthly chart of pretty much any market and not be able to define a trend as "generally moving up (or down) over time - making higher highs and higher lows".

    But since there are so many 'trends within a trend' depending on your timeframe traded, it is pretty hard to objectively test.
     

  27. This goes for everyone contributing here:

    Feel free to speak/debate about trends and their attributes. That is part of the reason I started this thread. At the same time, please do not argue and get off topic.



    As for the journal update:

    Two positions are still open, ESCL (+1.95%) NRPH (-0.12%)

    *percentages indicate change from my entering price. I am still using stops at -3% and taking profit at +3%
     

  28. thank you, runtrade.


    sure, one can see trends on a price chart. however, stating that higher highs and higher lows can be bought thereby increasing your odds of a succesful trade is not accurate. the odds of a percieved trend continuing after entering are consistent with randomness when tested.

    best,

    surfer
     

  29. How do you test someone's "perceived trend"? I believe the variables for all the different interpretations of trend are too numerous to test - this probably accounts for the randomness.
     

  30. let me know if I am understanding you correctly...

    If you enter an existing trend, the chances are random whether or not it will continue? Therefore, the only way statistically to be profitable from a trend is to enter before it begins?...which would mean that you would be trying to catch the knife and all data I have seen from that is that it is negative LT expectancy. How do you approach trading then?
     
  31. 1. yes.

    2. yes. but begining and ending of an illusionary, existing in the past construct is nebulus at best.

    3. no. actually, one's returns are increased following buying after consecutive days of market declines and decrease buying following multiple days of market gains. this is based on 15 years of data on SP 500 by larry connors and is outlined in detail in his book "how markets really work"

    4. i utilize mathematical models and fade extremes in sentiment. please see my public pre-katrina oil short, and my january GOOG short as examples of combining statistical modeling with fading extreme sentiment. i also tend to believe in mean reversion with stock indexes--- meaning declines are followed by rises and the opposite.

    best wishes,

    surfer
     
  32. Do you actually surf? Ive never seen a surfer come into a wave from the front, i imagine it'd be hard to do, sorta like shorting a "mean reverting" trend going from 90 bucks to 400.
     

  33. 1. yes, i have surfed.

    2. please note i am talking about stock indices--- not commodities, currencies, or individual stocks with my comments.,in most cases. on ocasion i do trade commodites or stocks with these ideas.

    best,

    surfer


    :)
     
  34. Interesting. You base your entries on a relationship with past prices (since that is how extremes are identified). And since your "mathematical models" implement past data in one form or another, then your independent coin toss earlier in this thread is an analogy that is completely without merit in the context of this discussion. You are talking from both sides of your mouth.

    And I'm going to say this slowly because you always seemed to miss the point in past exchanges: whoever tested the "validity" of trends did so on the basis of his own specific interpretation of a trend and his own working definition of what a profitable trend is. Therefore, his only legitimate negative conclusion can be that HIS interpretation of a workable trend did not successfully extract a meaningful edge from the market. He cannot legitimately claim that everyone else's definition cannot work. That is a generalization that is beyond the scope of what he tested. And please make no mistake, my definition of an exploitable trend is almost certainly different than yours.

    So remember, GIGO.
     
  35. A "mean reversion" can be characterized as a tendency, not unlike a trend. In fact, a "mean reversion" may be the "trend" that some people look for. Either way, it implies something other than the complete independence of past data, as suggested by your unfortunate coin toss analogy. Further, it illustrates the inanity of a singular definition of a workable trend for the purposes of drawing meaningless conclusions. It appears that your friend, VN, has a penchant for publishing "stuff" and raising ire. Not unlike you.
     
  36. Surfer seems to have clarified his approach as mainly applying to indexes.

    Not so much to commodities, currencies or individual stocks.

    To some extent I agree with him but I have found that if you look at a range of timeframes for different entities (indexes are almost the least trendy things around on the interday timescale) you can find opportunities to profit from detected trends. Similarly opportunities exist to profit from reversion to varying degrees.

    But, I'll be interested to see the thread owners effectiveness.
     
  37. His approach is his own, and I certainly have no issue with that. What I have an issue with is his regular and ongoing reference to bogus "studies" based on one-dimensional definitions that seek to reach generalized conclusions beyond the scope of the limited studies themselves. This is the basis for misinformation, and I think he should stop it. I don't know whether he is in over his head or if he just likes to draw attention for its own sake like his friend, VN. Either way, it is getting old.
     


  38. you assessment is seriously week.

    the sentiment extremes are not price related.

    what else don't you understand?

    surfer:confused:
     
  39. Really? No price relation whatsoever? Not even localized? When was the last time prices were at an all time historical low for a given market and the sentiment was "overbought?" Conversely, when was the last time prices were at an all time high for a given market and the sentiment was "oversold?"

    No, I don't think that sentiment exists in a vacuum.

    On the plus side, I find it amusing that you are talking down to me.
     
  40. Journal Update:

    Closed 195 ESCL @ $26.37 = +$150 Gross


    NRPH is still open @ $33.74 (+0.57%)

    I have decided to add a time stop of 4 days. After 4 full days, if the target is not hit, it is my opinion that the strength of the trend has declinded and should be sold since it is possible that the trend is ending (or reversing). Even with this, I can seperate the trades that hit either stop (+/- 3%) with trades that got stopped out because of the fading trend.

    TOTALS:

    Wins 2
    Losses 0

    Gross: +$360
     

  41. i am not talking down to you, whatever that means.

    i dont have any idea who you are.

    i understand what you are saying, but you are applying the wrong construct to understand what i mean.

    best,

    surfer


    ps. by the way, nice trading run !
     


  42. Only problem with this test is when the trend strengthens as with ESCL today. It closed at $28.47!! A whole 8% higher than where i took my profits. Will stick with the plan though...if anything, stop loss will be moved up to target profit if similar situations occurr again.
     
  43. I am really enjoying this thread run, thank you.

    If you are doing this just to show that TF in stocks can work you should keep you -3% +3% entries and exits, and don't add the 4 day rule.

    If you really want to create a trendfollowing system you should keep your 3% stop loss, but come up with a different exit. For instance, it could be let position run and if the close is lower than previous days close and position is still above 3% exit 20% of position. That way you can still have a position or part of a position on when ESCL is up 9% in a day. I don't think a time based exit works well, but may be necessary given your limited capital.

    Just MHO

    5yr
     
  44. I hope at some point you reconsider what you are doing and how you are doing it.

    The fact that you learned to reconsider your exits is a very good indication that you will be able to begin to think through each part of your current approach and replace it with a better alternative.

    Lets say one tenant you set up, the 3% exit target, is easily found to be a poor arbitrary setting. You could have known from the beginning that no stock of any sort follows a path that is as narrow as you selected.

    So you have just experienced selecting a stock that makes you about 5% a day for a few days. You chose to take 3% of that and then watch the rest of the run so far.

    I hope you go back over how to choose stocks too.

    Here is a for instance. Why not choose stocks that vary by X% and plan on making Y% of that run?

    You were looking at making 3% for the Y amount and you see Y is actually about 5% a day for a few days. The MAX in your journal name could be applied to Y and, when choosing stocks to the set X value as well.

    How will you fix the stop stuff you are doing? Lets say there are a dozen common approaches to setting stops. The ones that relate to MAX type thinking do not use an arbitrary Z%. Will you look at how the way you pick stocks relates to the way to set stops? I hope so.

    You picked a stock to make 3% and it fooled you and did what most high velocity money makers expect their stocks to do. What do you think made it fool you, speaking as a trend trader?

    Would you post your annotated graph so we can see the trend it is in?
     
  45. Journal Update:


    Opened 2 positions this morning...unfortunately, I got in about 5 minutes too late. I like to get in right after the bell. I lost a little ground because of that, but still confident.

    Long 740 TRE @ $6.75
    Long 150 RES @ $33.99
     
  46. Journal Update:

    Bad news, Res got stopped out half way through the day. Good news, I decided to add another $5k block to TRE and it paid off.

    Closed 150 RES @ $32.89 = -$165 Gross
    Closed 740 TRE @ 6.96 = + $155 Gross
    Closed 740 Tre @ 6.965 = + $159 Gross

    NRPH still open


    TOTALS:

    Wins 4
    Losses 1

    Gross: +$509
     
  47. Interesting thread. I hope it will spark more interesting posts.

    In the SFO Magazine's Jan 06 issue there's an article with an interesting take on the markets' randomness and predictability:

    "A new technology ... recognizes the markets for what they are: highly complex and usually independent of historical patterns, not necessarily random but marching to a different drummer of more fundamental and universal principles."
     
  48. Regarding trendiness dependency on the time frame (which in my experience seems true), this was posted by Alan Crary:

    "Just finished an analysis of the SP market. For the analysis I used the SP data from 1989 - 2001. Here's my results:

    The data from 1min. resolution per-bar through 30min. resolution per-bar is primarily in a trending mode. If planning on building a trend following system or intraday breakout, then these are the bar intervals I'd review.

    The data from 60min. resolution per-bar through 1 week resolution per-bar is primarily in a contra-trend mode. If planning on developing reversal systems, then these are the lengths I'd consider.

    The monthly data reverts back to trending mode, so if you need something to time long term investments (like mutual fund infusions), then this is the interval I'd use.(ex. 200 day MA).

    The data itself proved to be remarkably stable within the time frames. For example, the 30 min. bars had good trends in 11 of the 13 years. Likewise, the 60 min. bars had good contra-trends also in 11 of the 13 years.

    I did this analysis to get ready to build a really good daytrading system. Since daytrading is primarily concerned with the shorter bars, I'll have to look to building a system with the 1-30 min. bars using either a breakout or outright trending idea. "
     
  49. And this is another interesting observation about dealing with various time frames, that time frame analysis is dependent on the chart resolution not only its span:

    "In thinking out time frames it is necessary to understand that you cannot substitute a 10-period moving average of 1-hour bars with a 40-period moving average of 15-minute bars. Similarly, you cannot substitute a 10-week average with a 50-day average.

    It seems natural to think that any two trends covering the same time span will give the result, but that is not the case. Although, we can average many data points, we cannot get rid of all the noise; fewer data points over the same time span will always yield a smoother result. Therefore, the use of hourly, daily, and weekly time periods multiple time frames gives a much different picture of the market than simply using three different moving averages based on the same data.

    It is much easier to see the major trend using weekly data, find the short term direction based on daily data, and time short entry using hourly bars."
     
  50. Again, a position continued to increase signifigantly today. TRE ended +4% higher than my exit point. I am in the process of formulating the best way to gain this missed opportunity. I want to attempt to keep the conservative approach of "taking profits" since the whole idea behind trading with the trend is that I will have many more winners than losers, therefore I should not look for homeruns.
     

  51. actually, according to the book "trend f*llowing" by cov*l-- in which i am thanked in the preface--- LOL ! trend trading involves having many small losers to every big winner.

    best,

    surfer:)
     
  52. You need a good re-entry strategy if you want to cut your losses "shorter" without missing too much of the trend.

    I think you're on a good track, just I wouldn't be so mechanical about these percentages (3%, 4%, etc.) because it's like you want to squeeze the market into your expectations. You should try to adjust your expectations function of market's action.
     
  53. I heard it a lot, but lately even staunch trend followers as Van Tharp changed a little their position: he now recommends to close early a fraction (i.e. half) of your position for some smaller profit target and move your stops to breakeven as soon as possible.

    "Classical" trend following systems' large drawdowns could give you a heart attack, and really test your belief in your system's profitability.
    • Quote from marketsurfer:

      actually, according to the book "trend f*llowing" by cov*l-- in which i am thanked in the preface--- LOL ! trend trading involves having many losers and one or more big winners.

      best,

      surfer:)
     

  54. interesting. i did not realize tharp changed his views.

    thanks!
     
  55. I couldn't find a quote by Tharp, but the following is by another author in Tharp's newsletter (in the past this would've sounded like a heresy):

    "Back in the day, when trend-following was king and long- term position trading was the only method getting any press, using a profit target or anything other than a trailing stop was considered heresy. And for that type of system, it probably is. But many folks read stuff that was written for a largely trend following audience (including some of Van's early work) and make broad conclusions that profit taking exits are never useful. Questions about this show up on trading forums (including Van’s Mastermind forum) all the time.

    But more and more, I see good traders and system designers using hybrid exit strategies to optimize their results. Taking partial profits is not heresy if the exit strategy helps to capitalize on the strengths of the system design. Markets, even trending ones, have become more choppy than ever before. “Three steps up and two steps back” has become the normal operation in many markets.

    Trading professionals adjust for this by taking partial profits at pre-determined profit targets. They may let the rest of the position follow the trend. Or they might take the rest off at a higher pre-determined level. It all depends on whether the strategy is designed to capitalize on really big trends or not.

    This brings me back to the book draft that I’ve been reading. Another seasoned pro who has made consistent profits in the markets talks about his exit strategy. He takes 1/3 of his position off at a small gain, just to put the trade in the black. He then takes 1/3 off when the trade has made a bit more than he risked. The last one-third is used to trail a stop and try to stick with a longer term trend."
     
  56. Run,

    It sounds like you are getting away from your system, by adding on to a trade. If you are going to trade a system you should have set rules that tell you when/if you can add to a position.

    Also Grob made a very good point that stocks with different volatilies deserve stops that are different distances away. You saw this with the trade today. You could adjust your stops to a set equity at risk, 3% of total equity, and adjust position size and stop price to reflect the volatility. The higher the vol the smaller the position and the greater the distance to the stop.

    5yr