Home > Technical Topics > Strategy Development > Trading Winning Positions - The optimal approach

Trading Winning Positions - The optimal approach

  1. I was reading through a post on stops earlier and a few interesting things caught my eye. Specifically, this post:

    "I've done lots of tests on stops. On the tests for stops with a loss I found little difference in overall profits based on where a stop was placed. With tighter stops the percent of winners dropped but the winners made way more money than the losers. With looser stops the percent of winners went up but the total win/loss ratio decreased so the overall profits were the same. I've found it much more profitable to spend my time on the stops for profit and I think it's a big difference between ok and better traders.

    One area of research you might want to check is entry size. If you don't do all in at once it can have a big difference in your trading.
    I found if I go in partially initially, and then after a fixed period adding more if the position is profitable, the overall profits went up, the time between equity highs went down, the drawdown increased, and the sharpe ratio went down. If I did partial entry at first and then adding if the stop hasn't been hit, but the position is underwater then adding to the position gave lower overall profits, higher time between equity highs, lower drawdowns, and increased sharpe ratio."

    Lately, I have been more focused on improving my trading, and an area I think I could improve upon is getting more out of my winning trades. Then I figured that this is an area we all could do improve in, and the focus of this thread.

    As I see it there are two ways to improve the way we trade our winners:

    1. What is the optimal frequency and/or strategy for
    adding to winning positions?

    2. What is the best overall stop strategy once you are in
    the position?

    I look forward to your posts, and good trading to all!

    Mike

    P.S. I will be posting my research and strategy a little later in the week.
     
  2. I read Jesse Livermore book once and he always added to winners, thus his losers were small, and winners big. But he used higher highs and higher lows when buying, and added on specific breakouts. Thus, if he was wrong on adding, he figured he was wrong or ready to get out completely. eg. stock makes a "lower low". In this way he would exit his entire position.

    This seems like the best strategy to me. It is more active, and requires more analysis.
     
  3. You're definitely on the right track with your experiments. As far as your questions go, what works for me is... if you do decide to add to winning positions, buy on 'secondary buy signals', usually dips on low volume etc. Usually I enter with my full position and if I get stopped out I get stopped out. My position size is based on how much I'm willing to risk per trade. As far as exiting winning positions and stops, I've found that a nice trailing stop works well. I like to set it below support levels as the stock moves up.. or sometimes in a certain %tage to lock in gains. Knowing when to exit winners is a pretty hard thing to master, because I guess the inherent greed 'maybe it will go up more' is there. So the trailing stops help lock in a part of it, and usually if you see the position about to turn sour (hitting resistance and not able to pass it, reversal formations or indicators) it doesn't hurt to exit half of your position, with the idea that you could always get back in at any time.
     
  4. What I currently do is

    full size on first position

    then 1/2 size a few cents up.

    So lets say


    STock is 37.20

    I buy 2000 shares 37.20

    when it goes up to 37.25

    I buy 1000 more shares

    if it falls back to 37.23

    I sell 1000 shares

    If it goes up to 37.26 , leave it alone.

    Of course with this you need to micromanage, but There is the lowest risk @ initial entry , and the continuity is usually good.


    So I'm developing a better strategy where I layer it 3 times,

    1x. entry
    2x. 3-5 cents after
    3x. 15 cents after?


    I think with this kind of stepping, I Can build up a nice 3000-8000 share position without any risk depending on liquidity.
     
  5. The basic rule for adding to a position is to only add when the addition doesn't increase your initial risk.

    Example

    Say you use a trailing stop of 10 pts. in the ES.

    In you're initial position you short 10 ES at 1289.
    Your risk is then 10pts. * 10 contracts * $50/per-contract = $5,000

    Now say the market moves to 1287. The trailing stop now only has 8 points of risk in the overall postion. 8pts. * 10 contracts * $50/contract = $4,000 overall risk. You could short an additional 2 contracts without increasing risk beyond the initial point. 8pts. * 12 contracts * $50/contract = $4,800.

    This is how I was taught to add to a position while keeping the risk constant.
     
  6. A lower low or higher high sounds easy. You have to get a few bars along from each one before it can be identified in which case you can end up saying, I've missed it.
     
  7. This is similar to what i use and it works well. Isn't this what Phantom of the Pits recommends as well??
     

  8. I've made it a point to study the same posts you're referring to and can lend a little insight. Most breakout and some trend following strategies enter using tight stops because anything more would 'be a waste', to paraphrase one poster.

    The inverse rationale behind the, 'anything more would be a waste' comment, is elucidated by other members which can be summarized as, 'once entries are made, trades go profitable and never look back'.

    Again, 2 sides of the same coin. Just traders expressing an entry phenomenon - usually for breakouts or trends - a different way.

    Both perspectives serve to reinforce the superiority of tight stops for some strategies. Again, they don't work well for all.



    Again, more sage advice. You need to learn when to step on the gas. Increasing the size and frequency of winners is just as critical to minimizing drawdowns as the direct approach to neutralizing dd (usually money management).

    Pyramiding or stacking is really important and can be incorporated into a winning strategy to really press winners. Personally, I stack. Depending on the strategy, winning entries that occur in runs (dependency) enjoy a higher win probability after each iterative succession up until a certain point. Therefore, under such conditions, pyramiding or stacking makes statistical sense.




    Ultimately, optimum position sizing needs to be tailored to a specific strategy.
     
  9. Now that he's here, acrary and Steve46 posted a goldmine of information in regards to the importance of runs or 'streakiness of wins'.

    My reason for highlighting their contribution is high dependency between winners (runs) offers the most reliable foundation for which to pyramid, IMO; as pyramiding is essential to keeping dd low and total wins high.

    So thank you acrary and Steve46!
     
  10. People are pretty tight lipped on this site, aren't they?
     
  11. Hello:

    I think a couple of things should be said

    I notice that most retail traders direct their attention to setups and entries.

    Commentary about stops, position sizing, position management and exits is almost never the primary subject of a thread here.

    Ironically, these issues are significant and interestingly, people seem to have little to say. You comment that people here are "tight lipped". I think not. I believe that most ET members are unable to offer anything of substance on the subject.

    As I move on in time, I have tended to go believe that widers stops are more effective for daytrading. I found long ago, that knowing what the best stop size is, and trading it are two different things.

    As regards adding to a profitable position. I was taught the same strategy as Acrary. In addition, I try to incorporate a couple of ideas as follows;

    1. Add when price moves past a significant resistance (pivot, or resistance).
    2. When you have a profitable entry, add on the first pullback

    Finally, I add based on an analysis of my market. What I mean is for the Russell contract, I know from experience what the average length of line is for each intraday swing. I will only add early on, so that I can take as much profit as possible when it works out. If I have a significant profit, and I am close to the point where normally I might close out, I will not risk it. Instead I will exit all but a small piece just in case it continues to run.

    My overall philosophy is to make sure that my average winner is much bigger than my average loser.

    I hope that helps.

    Steve
     
  12. My post on adding to a winner is only valid if you adjust your stop as the trade goes in your favor. But what if the system you develop uses a fixed stop that doesn't change during the trade? In other words I'm putting on this trade with my stop. I'm holding until I'm either stopped out or I get a exit signal from my system.

    If that's the case you can do tests with adding to a position to see how good your entry is. To evaluate improvement I'm going to look at the profit factor of the system. (I made a case for judging a system by profit factor and trading frequency in my system devel. thread.) I know if I increase the PF while keeping all else equal, I'll make more money and more consistent performance.

    I'll work through a example system to show what I mean about using add-on trades as a tool.

    Here's my old discard system dsc389 used on the NQ market from 2002-2005 to use as a baseline. Notice it's using 10 contracts per trade.
     
  13. Before I get going I wanted to address a pm I got on why not just add to a position when it's underwater if the overall method is profitable. The old dollar cost averaging approach to trading.

    Here's a add to the base system when the position is underwater. From it, you can see it does nothing good for the system. The profits, profit factor, drawdown, and % winners are all going the wrong way. This is why you don't add to a position when it's underwater.
     
  14. Here I'm adding a small amount of 2 contracts to the position when it's winning by a certain amount. You can see it improves the profit factor a little bit. Since I know the stop is fixed this tells me the second entry was better than the initial entry.
     
  15. To show that the second entry is better than the first I'll add the same number of contracts as I did in the initial entry. This shows the improvement in profit factor more clearly.
     
  16. With the knowledge that the second entry is better than the initial entry I could just enter one contract initially and then the other 11 with the second add. By doing so you get to see the full effect of using the second entry instead of the first.
     
  17. In real trading I could just leave the system as is and use the first entry as a signal that the second is coming up. It'd be easy to frontrun the second signal and boost the overall trading to a PF above 2 using a little discretionary trading.

    Anyway, I just wanted to add that you can try using the addition to a winner process to see if you can improve your initial entry. In this case you could see a pretty good improvement.
     
  18. I also find it very important to know where in the lifes trend you are.

    I prefer tighter stops later in the trend and looser stops in the beginning.
     
  19. Thanks for sharing, acrary. Can we assume your posted example exhibits dependency between trades?
     
  20. I am getting different results that suggest building into a position as a trade goes against you is better. Perhaps this rule is system dependent?

    Using a contrarian strategy for range bound markets (sell on local maximums, buy on local minimums) I get the following results:

    Buy/Sell 3 contracts at trade initiation:

    Net Profit: $587,100
    Max Draw Down: $68,070
    Net Profit / Max Draw Down: 8.62
    Profit Factor: 4.74
    Number of Trades: 66
    Periods Profitable: 61%


    Buy/Sell 1 Contact at trade initiation and as trades enters a draw down. Max 3 contracts held:

    Net Profit: $545,990
    Max Draw Down: $59,160
    Net Profit / Max Draw Down: 9.23
    Profit Factor: 5.60
    Number of Trades: 172
    Periods Profitable: 68%

    My profit factor, number of trades and number of profitable periods increases when I build into positions during draw downs.

    I will also perform this test on a trending strategy.

    Granville
     
  21. When applying it to a trend following system:

    Entering 2 Contracts at trade initiation:
    Net Profit: $205,440
    Max DD: $29,640
    Net Profit/Max DD: 6.93
    Profit Factor: 2.95
    # Trades: 66
    % Periods Profitable: 61%

    Entering 1 contract at trade initiation, 1 when trade becomes profitable:
    Net Profit: $194,550
    Max DD: $28,960
    Net Profit/Max DD: 6.72
    Profit Factor: 3.01
    # Trades: 126
    % Periods Profitable: 61%

    Entering 1 contract at trade initiation, 1 when trade enters draw down:
    Net Profit: $199,390
    Max DD: $27,380
    Net Profit/Max DD: 7.28
    Profit Factor: 3.21
    # Trades: 125
    % Periods Profitable: 64%

    Notes:
    1. These two systems are always in the market and are long term in nature - not day trading.
    2. They may or may not have an edge, although I suspect from some statistical tests run on the range bounded system as described by Acrary, that the first one has an edge.

    Observations:
    1. On this example, doubling the number of trades and increasing the profit factor did not improve my % of profitable periods. This may be because I am taking the 'Global Profit Factor' ove the life of the system rather than on a period - by - period basis.

    I am suspecting there is something I am missing.

    2. Entering added positions when in a draw down did improve the risk adjusted results.

    This is against conventional wisdom and other results posted on this thread.

    What am I missing?
    What could be a faulty assumption that I am adhering to?

    Granville



    1.
     
  22. Acrary,

    Great stuff. I was wondering if you can share your views on systems that have a stable profit factor but unstable expectancy. Say for example a system has a stable profit factor year after year but the expectancy is declining. The number of trades is more or less the same every year.

    If I understand correctly, the stable profit factor will keep the system consistently profitable, but the actual $ profits will decline each year (assuming no money management for now).




     
  23. Hello,

    I am relatively new to trading. Well, my advantage is that I lost (and learned!) a lot in 2000 and stopped until this year. I am more careful, objective and disciplined now and determined to make a comeback. Did a lot of reading and all that. Liked Livermore the most although in my opinion he took big risks.

    No daytrading and I let the market decide how long to hold. I trade long on high volume breakout after stock has been consolidating (downs must have low volume) but ma200 is going up significantly and the general market is positive such as right now. I buy at market about one percent above resistance which then becomes support. I do not wait for pullback because some of the most profitable stocks do not pullback and if there's a pullback usually it's minor. I buy one third of my intended maximum. I like to have a good start because then its almost sure I will not lose money which is the main objective. Usually trading on breakout I have few percent profit the same day already. On the next days there may be a pullback until buying price but I don't care.

    My stoploss depends on the resistance and the maximum I want to lose. I don't want to lose more than 1% of total capital. Also I always place stoploss a few percent under resistance which has become support after breakout. My experience is that stoploss is almost never touched trading on breakout and it can be tight, 5% or so but I do not hesitate to place it at 3% if possible.

    While finding the optimal approach I used to trade not on breakout but when stock was going down and near to ma200 with ma200 going up. Results are not as good and sometimes stoploss was hit. This is not the way to select winners I found.

    I double up after first trade has shown a good profit and after next breakout from consolidation. Very important to have a good profit because if price goes down after second trade I lose money quickly. Then I repeat the same trick. So I have two stoploss levels. One for the first trade and one for the second trade. If I am stopped out, I will still have a good profit from the first trade. If it looks good: price does not fall back to second stoploss and second trade shows a profit, I move the first stoploss to the level of the second.

    This I repeat one more time.

    While stock goes up I widen the stoploss so I will survive a natural drawback of 15 or even 20%.

    It has been working well on last 25 trades or so but I am still building up experience like everyone.

    regards,
    Ivo
     
  24. Jessie committed suicide and died dead broke and deep in debt.

    ..but maybe that was for reasons other than his stellar trade management.
     
  25. Nobody gets out of life alive, so what if he shot himself at the age of 63, he had already lived beyond the avg male life span for that era. He is still the greatest speculator/trader of all times. He left behind $5M, in trusts and real-estate - inflation adjusted for 2017 that is $88M.

    Livermore is the greatest trader of all times period. He had been up and down a few times - the thing is nobody has ever been up as much he was. He bet big & lived big, he went out with a bang. What do you expect - some tight wad like Buffet that worries about how much he spent on his Egg Mc Muffin?
     
  26. I agree with you. The $5m he left behind makes him a big winner despite the suicide. $5m during his time was a huge huge sum of money. His suicide had less to do with his trading performance. It had more to do with mental depression. I would say his later years of trading which turn out to be poor were probably related to depression. The best thing he could have done would be to stop trading, given his unfit state of mind. Anyone who feels depressed with tendency of suicide despite having $5m is highly suspected to suffer from some form of mental illness.
     
  27. Mr. Gumby, that is very deep. WTF man. I am attempting to process that ideation. I am failing at this task. I do not like you at this current time.

    Flom.
     
  28. Only add to winners if you have a different entry point that is more profitable than the first, and if that's the case, don't bother adding on the first one.

    if you add to winners and price goes against you, you will lose more. If you think price is going to go up, why not go long with your entire position other than in cases where:

    If you add to a loser a small movement makes it easier to make money. Presuming you are sure of your initial direction, going all in at once or adding to losers makes the most sense.
     

  29. The standard advice on adding to a profitable position is to wait for a new valid entry signal and enter there. Traders are usually advised to move their stop on the first trade to at least b/e and possibly reduce it at the same time.

    I go against these rules. I always add a duplicate trade to every winning trade as soon as the first trade reaches b/e. I don't wait for a new entry signal - rising price is the entry signal. If b/e is say 87pts above entry on a long trade, then I keep adding a new trade every 87pts up and moving all stops 87pts higher. I never reduce a winning trade. The maximum capital risk trading this way can never be more than 87pts' worth.

    This is contra most rules on pyramiding and if you're going to try this demo it first.
     
  30. B/e means break even? If so what do u mean when u add at b/e when it is above entry? Surely your entry is b/e?
     

  31. Sorry to cause confusion - yes, b/e means break-even but I meant to say price.

    A second buy order would trigger at first trade's entry plus 87pts, then the first trade's stop is moved to entry, which is b/e. A new trade opens every 87pts higher and all previous stops are raised 87pts at the same time.
     
  32. I think I do not get it upon reading your recent post.
     

  33. Its a simple approach but easier to do than explain. The idea is to have multiple trades opened at the same interval as the trend continues. Each trade has the same risk in £. Only the last trade loses £ capital, the one before it breaks even and all the others lose each the same risk in £ but that is from unrealised gain. The net gain realised when all trades close simultaneously is far greater than the losses and far better than holding a single with-trend trade throughout the same distance.
     
  34. Are all trades long or short? Or are you doing both at the same time? Are you saying you have more evidence the trend will continue as it goes, otherwise why not not just open with a larger position?
     
  35. "some tight wad like Buffet that worries about how much he spent on his Egg Mc Muffin?

    Now, that's funny.
     
  36. My read is they're all long or all short trades. Using numbers, you buy at 100, stop at 99.13. If the price moves to 100.87 then you buy a second unit and your old entry (100) now becomes the stop. The loss on your first unit would be zero because you bought and sold at 100. Your loss on the second unit would be 87 pts.

    If you bought the second unit at 100.87 and the price continues up another 87 points to 101.74 then you buy a third unit your new stop becomes 100.87. If the market hits that stop then you've lost 87 pts on the third unit, nothing on the second and made 87 on the first unit. Again the total loss at any point in the trade is always going to be 87 pts.

    The advantages are it's a simple rule which is easy to keep track of and code in need, secondly it gives a safe way to build a big position.

    If you get up to a 10 unit position for a cost of only 87 pts then that's much lower risk and higher return than putting on 10 units straight away with an 87 point stop which costs you 870 points and you're not guaranteed any profit.

    If you've got a 10 contract position using the tomorton process described above, then you would already have 8 guaranteed in the money contracts, one at money and one out of the money. Also by implication the first position would be in the money by 8x the stop distance (if my mental counting happens to be correct!).
     
  37. That's great Cat88, clearly put.

    I don't use r:r but this grid system quickly breaks into double-digit r:r. I don't know of any other strategy that even considers double digit r:r as a possible outcome.
     
  38. There's a second more subtle piece of knowledge which sounds obvious but reveals the insight of the process, which is that you cannot know which of the initial moves is going to become the big winner.

    With hindsight we can say that was a big winner I should have put on 50 units! But this is actually death unless you are purely mechanical and have rigorous statistical support for your bet sizing. I.e. you know how many times a trade set up will become a disproportionately big winner.

    What it really points towards for me is that as a winner continues to win it's survivability rate actually climbs, so a move that's gone 5% in one direction is actually a lot more likely (say 15%) to continue in the same direction than a trade thats gone 1% in your favour (which is probably just white noise).

    A move that's 10% in your favour has a much higher probability of continuing in your favour than reversing because so few trades actually go that far. Maybe only 3 trades will have moved 15% historically so your chance of becoming one of 4 all time 15% is 25% which is better than 15% chance, which better than the white noise.

    It's like every hall of famer a sport person passes, the more likely they are to be the greatest of all time (as measured by stats). Hence add to your winners but don't spend a 100 million bucks on a college prospect.
     
  39. Thanks tomorton - but what does r:r stand for? risk return?
     
  40. Nearly, its risk:reward, same thing.
     

  41. Love it, this is exactly the point of pyramiding in a trend-following trade.

    I use the length and consistency ("smoothness") of the trend as optional criteria and the points score these contribute to gives me preference towards those charts. Using the straightforward % price change in x days is often a misleading statistic.
     
  42. This doesn't sound like you are very confident in your positions.

    Enter, if it goes in my direction keep adding. Picture you adding at 20 with an exit at 19, it goes to 21 and you buy one more. If you move your stop up to 20 your risk is now 2, not 1. Here, an equal move against you will have more loss than you would originally. You have to move the exit to 20.5 in this case.

    The math that you mentioned is wrong.
     
  43. Then if you add another one at 22 you have to move the stop to under 1 away.
     
  44. if i bought at 20 and exit at 20 then there is no loss aside from transaction costs. If I bought one at 21 and exit at 20, then the loss is 1 unit.

    looking at it another way - when the price got up to 21, the stop is moved to 20, so you have a unit at bought at 20 and a unit bought at 21.

    The average price is 20.5, if you exit 2 units at 20 then 0.5 * 2 = 1 unit loss.
     
  45. Also whether you are 'confident' or not in your positions is totally down to your personality and system.

    A mechanical trader (which this approach sounds like it is) would say being confident is irrelevant because they are simply following a procedure which is expected to get stopped out at some point (either profitably or not).

    A discretionary trader would need to be confident because they rely on their superior judgement but you would need to know the person is a discretionary trader to assume that confidence is meaningful to them.

    Also confidence in a 'position' is an interesting choice of words. I would say confidence in your own judgement is more important because confidence in a position suggests you know it will be a winner and that's why you are going to hold on. Whereas confidence in your judgement in your ability would make it easier to say "that was a bad call, I'm going to walk away" because you no longer have confidence in your position (but you do have confidence in your judgement to know that this particular trade is not going the way you envisioned it might).
     

  46. No. For one thing, no judgement is required with regards a particular trade. If the chart set-up matches the pre-determined parameters for a trend-following buy, then its a buy. I don't trouble over entry patterns as long as not buying at a new high. All trades are of equal stake. All trades have a pyramid order set as soon as opened, no judgement is required on this.

    Secondly, a new trade is only added when the SL on the one before can be moved higher to balance the new risk. So, as Cat88 says -
    buy Trade 1 at 20 with SL at 19: account capital risk = -1
    buy Trade 2 at 21 with SL at 20, move Trade 1 SL to 20: total account capital risk = -1
    buy Trade 3 at 22 with SL at 21, move SL Trade 1 to 21, move SL Trade 2 to 21: total account capital risk = 0

    From this point there is never any account capital risk arising from additional trades but the profit side increases parabolically.
     
  47. Your math above is wrong.

    Buy 1 at 20, SL 19, loss 1.
    Buy 1 at 21, you are now long 2 at 20.5 and your stop has to be 20.
    Buy 1 at 22, and you are now long 3 at 21. Your stop has to be 20 2/3.
     

  48. This is a great objection, thank you for the thought you've put into this. Most teachers of trading would go along with your view, accepting that a major objective is to use the SL to protect the unrealised gains: but that isn't a primary objective for me.

    The key point of departure is in how you regard the trades. The idea is to treat added positions as NOT being aggregated. So you treat it not as a progressively larger and larger position, which needs an increasingly tighter SL, but always an increasing number of parallel positions of equal size.

    So, the distance in points / pips from last entry to my SL is never changed. This does lead to significant loss of unrealised gains when my new SL is hit, but it also means that the SL distance does not get progressively narrower and therefore increasingly likely to be hit before the trend gets to a really serious correction. As in your example, the SL is already getting considerably narrower after just 3 trades: after 6 or 7 it would be likely to be hit by just mild volatility.
     
  49. Exactly why I don't add to winning trades.

    Listen, you're talking to a guy who understands that being long and short at the same time is not the same as having a smaller position open in one direction. I get that even though the math is the same, they're not the same. I get that not combining all your positions into one is different than how I did it, even though they calculate out the same.

    I cannot imagine a system in which you keep adding when it's going in your favor. Buy at 50, it goes to 51 so you buy another, it goes to 52 so you buy another. Really? Do you just treat them separately? Contract is now down to 51, what do you do? You're net even, +1, +0, -1.
     

  50. Yes that's right, you treat all the trades as separate - my platform allows me to display them as separate trades but other platforms might need a bit of record-keeping. But this can be done in advance once you know the distance to the SL from the entry of the first trade - after that, the gap is always the same.

    The objective is to increase the aggregated position size but without increasing account capital risked. However, I know the account capital maximum is at risk for longer and the unrealised gains are at risk of significant top-slicing. But I find and feel - and more importantly calculate - that the parabolic increase in profits is worth these downsides.
     
  51. What number of buys do you do before you say right, I think it's done going up?
     
  52. I think the point is that you don't presume to know when the trend will end. That's the whole point of following the system in a mechanical manner. Some you might get stopped out on the first unit, some you might get up to 5 contracts and otherwise rarely maybe 20 contracts. We don't know what the market is going to look like this time so just follow it as far as it goes. Obviously longer uninterrupted runs are rarer than shorter ones but how can you know when copper or OJ or cocoa or 10 yr notes is going to have a 10% or 20% move with limited pull backs? You don't. But that's a big payday to miss out on because you 'thought'. Mechanical systems like this are a way to ensure consistency and answer questions such as when do I take profits off the table? Or an alternative criteria of answering: is the momentum in this move done?

    You sound like a discretionary trader so if you put a trade on, how do you decide when to take off a trade? And has this been successful?

    If so then great, if not then try another criteria. Tomorton is just suggesting another lens through which to view the same thing (when to buy and when to sell).
     

  53. There's no limit. So if the trend continues without serious pull-backs sufficient to trigger the universal stop, you keep adding to it. The important thing to bear in mind that I only regard risk as risk to account capital: as this can never increase above 1.0 and as after 3 parallel trades it has fallen to zero, I have no reason not to add another and another position: and as the trend continues strongly, that gives the opportunity.

    As far as I have ever heard, my pyramid tactic is the only one which has even a possibility of reaching r:r in double digits, which I believe it achieves after 7 parallel trades.
     
  54. Can i ask tomorton what sort of time frame you typically trade? Are you trading based on daily data by which I mean the open high low and close for one day comprises one entry on the chart rather than intraday time frames ?

    Also have you found that using some sort of weighted volatility measure skewed towards recent data is better than anequally weighted volatility measure? For example you set the stops for a long position at the entry price minus the the average true range for the last 20 days (or some other multiple of recent range).
     

  55. I trade off the dailies only. I take account of weekly bars only in comparing strengths of trends on different charts.

    It could be. If I can't find a TA-based SL level, say the bottom of a swing low for a long, I would use 2 x ATR14 to set the initial SL. But if this works out at say 83pts, then I retain the 83pts as the "grid" right through this series of trades on this market.
     
  56. Only thing I would add is some "escape plans". If you're long, exit manually if price closes below the 50EMA or if the 20EMA closes below the 50. If you're long equities, exit manually if the Dow prints either of these.

    As I've posted previously, these rules would have got you into cash before 16 of the 20 worst 1-day falls in the Dow since 1900.
     
  57. Here. Price is 50. You add 1. Now it goes to 51. You add 1. Now it goes to 52. You add 1. Now it goes to 51. Last position closed. Now it goes back to 52. Do you add another one? Now it goes back to 51. You have a number of losses. Now it goes to 52. Do you add another one? Now it goes to 53.

    To what extent do you let it go back and forth in the middle?
     
  58. Considering MA crossovers do not have forecasting uses, why would you exist a long position when a 20 crosses a 50?
     

  59. No. A new position is added at each grid increment, i.e. 1. All stops on all positions are set at the penultimate grid price. So if price goes to 52, all stops on this latest trade plus all already open trades are set at 51. When price hits 51, all trades close simultaneously.

    The chart now goes back on the watch-list in case the TA would support a new trend-following buy and the start of a new series of pyramids.
     

  60. I would consider a trend-following long if as many as possible of my TA criteria are met, but definitely the three mandatory criteria -
    1) the 20EMA is above the 50
    2 the 50EMA is sloping upwards
    3) the 50EMA is above the 200

    If already long and one of these three criteria is eliminated by price action, it doesn't make a lot of sense for me to continue holding and adding. I could live with a little negative slope by the 50, but when the 20 crosses below, that would just get me out every time. Though it would be rare for the 20 to cross below the 50 without the universal stop having been already triggered.
     

  61. You should take the opportunity of suggesting a better way forward, at least because it might help me improve what I'm doing.

    Let's suppose price is at 50 and already in a consistent uptrend. We enter long at 50 and the TA conveniently suggests a SL at 49. Let's suppose price continues upwards to 56 without any significant correction but then drops to 55.

    What I would do is buy at 50, 51, 52, 53, 54, 55 and 56. The SL on all open trades is always set at 1 below the highest new entry price. When the 56 trade opens, all the stops are at 55. Now price drops back to 55, and all trades close simultaneously.

    My net profit from the 7 trades is 5 + 4 + 3 + 2 +1 +0 -1 = 14. That's 14pts from a maximum move of 6, and a stopped out price of +5, and achieved with no account capital risk greater than 1 at any point. If these things mean anything, that's a r:r of 1:14.

    Please go ahead and outline an alternative strategy given the same uptrend scenario that matches this without greater risk and with at least equivalent r:r.
     
  62. Win ratio is more important....than any other ROFLMAO
     
  63. Me too.
     
  64. Small pull back trend day. Great day for adding to losing positions in the ES. Over and over again.
     
  65. Grinding higher and higher.....someone said that would probally happen....
     
  66. Of all the metrics win rate trumps all.
     
  67. Trumps a winner and win rate trumps all! ROFLMAO
     
  68. Now this isn't the same as forex grid trading?

    Here buy one at 50, price hits 51 and you buy one and move all stops to 50. Price hits 52 and you buy one and move all stops to 51. Price hits 51 and you close all positions.

    Can I consider this correct?
     
  69. Provided you were uncertain that the trend was going to keep going up, this way makes sense. Otherwise you would go long with your largest position first and make 35 instead of 14. This sounds like you are not able to successfully predict market direction. Do not feel bad regarding this. You are not the only one. Hence this forum.

    Risk reward doesn't mean much, though. When price touches 50 I can buy with a stop loss of 1 and a goal of 80 and have 1:30 risk reward. Who cares?
     

  70. This is what you do but I don't know if its the same as forex grid trading.
     

  71. Well, if you know which uptrends reaching 50 will continue to 56 before they suffer their first 1pt pull-back, I wish you good fortune. I am sure we have nothing more to discuss as I am simply not able to trade on your level.
     
  72. @tomorton I have used the exact same idea years ago and then discarded it

    upload_2018-2-16_4-21-14.png

    as you can see on paper it looks very good ...but you have to capture 2 grids before becoming profitable..

    As the entry condition does not matter the system depends on continuous price trend without retracement by 1g.

    the problem is price action very frequently moves/retraces by 1g regardless of what you choose as g.

    After a sample number of trades

    what I observe is roughly 45% trades are -1g and 25% trades are 0g
    so 70% trades end up losers

    And 10% trades are 1g which returns 0

    rest 20% trades will capture 2g to 7g ...anything beyond is very rare.
    and with each additional g the % number of trades will reduce...
     

  73. Nice post, thank you.

    Yes I know that though the loss of -1g can never be exceeded, it persists across 2 trades, so twice the holding time so twice the possibility of being stopped out.

    What I'm interested in is trading off the certainty of never having a greater loss per trade against the possibility of having parabolically greater wins per series of trades for no additional capital risk.

    We shall see how this works out of course. Its possible that your results are less advantageous than mine will be, we will see.

    But surely you must have realised that if only 5% of your trading examples made +3g, while all the remaining 15% of winners made only +2g, and no trades at all did as well as +4g, even your figures indicate break-even? So I think this approach bears exploring and don't think I am taking an excessive risk.
     
  74. Is this a new idea you are working with ? or
    Is this something that you have been working for a while ? if yes how is your result so far?
    If I remember correctly yes my result was breakevenish for this.
    The idea definitely had some potential
    wish you luck on cracking this one.
     

  75. This is quite new to me. I ran my first pyramids using this grid approach from late October after about a month of paper trading it. Since then I add a pyramid entry order to every trade that I open. My trades are trend-following only so obviously they tend to move in the trend direction more than 50% of the time and they often don't have a "natural" limit of travel, like a H&S or Triangle etc. would.

    Results so far are not numerous because of my trend criteria, and I'd like to run more to be fully confident. But January was my best ever month in trading since 1999: two-thirds of all wins and nearly all the biggest wins were pyramids. I'm not taking any more capital risk than I would on a single stand-alone position off the same chart so I can't see a reason to not press on. Though capital risk remains in the early days, it never increases, though the aggregated position can be built up to way more than I would otherwise dare.
     
  76. In so many words, a low win rate? Win rate is the most important metric ...Imo
     

  77. Some say win rate is the most important, some say r:r. I don't log either but both matter: my strategy is inherently biased towards both rising.
     
  78. IMO second most important metric is average win vs average loss.

    In your strategy an instrument that tends to have prolonged trends could work well but such an instrument will also work well in terms of scaling into losing positions on pb’s as one is buying cheaper (in a bull trend). The problem is trading instruments that have alot of reversion to the mean.
     
  79. Here is a twist. Have you tried using the methodolgy you are describing ON an instrument that reverts back to the mean alot but using it scaling into losing positions? That is, use it as a method to scale in. Scaling in by the grid!

    I generally scale in by steps with each step being at least the min scalp i would want to play in an instrument. If volatility is high that day then that scale in step naturally becomes bigger. In other words, the scale in point is dynamic and differs according to the price action of the day.
     

  80. Not deliberately. I only take trend-following trades, that's what this grid/pyramid system is built to capitalise on. So a trade might unintentionally be on a mean-reverting instrument (probably in the wrong direction) but I'm blind to this characteristic when selecting the next target. I look at S/R when trying to find a stop-loss price and this could coincidentally line up with the limit of travel on a mean-reverting instrument, but I ignore all forms of resistance in uptrends and support in downtrends.

    Apart from which, mean-reversion works until it stops, and surely you only know its stopped mean reverting when it doesn't, by which time you're already in the trade?

    I don't see scaling in gives optimum return. I want to be in the trade up to my capital risk limit right from the start or else I'm in cash.
     
  81. By Forex grid trading I mean opening one at 30, opening one at 40, opening one at 50, etc., and rarely closing them, and selling lots in the reverse direction too so you get a bunch that are open all at once with the idea that as price increases and decreases you make a lot of money over time as trades close.
     
  82. Do you think these would no longer be problems if you started with 2, then added 1, then added 2 and proceeded that way? That way your stop out at +1 would still be a winner but not break even.

    Having it this way requires good entries.
     
  83. Do I think we could do it with scaling and adding to losing positions? Probably.

    What about buy some, it goes against you, buy more, then it goes in your favor, buy more.

    I see one concern with this. You have to be sure about identifying bottoms.
     

  84. I only open and only add to trades that are with the trend, if there is one. I don't have as much faith in top-picking and bottom-picking as I do in recognition of an uptrend and a downtrend. I also don't put as much faith in a young trend that's just come off a reversal and inherently has no TA characteristics of its own as I do in an established trend that has printed a smooth consistent progression.
     
  85. I understand. The comment I made was based on bpr's intelligent post and I'm trying to uncover a way to make this better. This way works if you can pick trends that aren't going to drop on their way up. I rarely see a trend that does that. What about just using OTM options instead? Same idea, low cost that will result in a loss or a big win if it keeps going up potentially.
     

  86. I've been a trend-follower for several years but found that there were two reasons why I started losing money a little while back -

    1) I started being selective about only taking strong trends: I make money when I take every trend that fits my basic criteria. I don't know of a way to pick trends that will be consistent and smooth in the future.

    2) I cut down my initial entry signals to a very few very closely defined candlestick patterns: this meant I was late getting into the "strong" trends or missed them completely: I now take any price that's not a new high in an uptrend / new low in a downtrend.
     
  87. Ironfist determined similar conclusions in this thread, legitimately that you have to take every opening to make money in a trend following system. Additionally, that's the only thread I've ever seen someone give a complete definition of a trend and how to identify it in real time. I ponder if you could use your entries and his guidelines to make an entire system.
     
  88. With proper management one can enter on about 95% of the bars. At anytime. On any bar and make money. Every bar has bulls and bears playing it all day long. Management trumps entry almost all the time. Most traders focus their concentration on entries. What is far more important is context and management. I too use setups and context for entries as there can be a slight edge there, however, if it goes against me soon after entry I focus more on context and the behaviour of the adverse move to determine if I exit with a loss immediately or add to the position, which at that moment is a losing trade.

    There are no perfect entries when taking one. Never have been and never will be although in “hindsight” we sometimes see perfect entries..mostly luck...most definitely not skill or we would execute them nigh everytime.

    Context...management...win rate...average win bigger than average loss...these are the important issues to me, at least. Context and management implies the ability to read PA well enough to achieve the above two metrics consistently.
     
  89. This sounds like you're talking about money management being the factor here. Are you holding when it goes in your direction for many bars?
     
  90. Both Risk and PT management. As to how many bars I would hold before taking profits, it is, never (at least for me) set in stone (although I will set an estimated initial PT) but the actual PT and even the SL are both subject to the dynamics of the market in the context in which I place the trade. In other words, I do not say I gotta have a 2:1 reward to risk or 4:1. If market dynamics after entering the tells me i can probally go for. 6:1 I’ll wait for more bars and more profits. If, on the other hand, my initial target may be 3:1 but after entry the subsequent dynamics tell me I will be doing good to get 1:1 then I do not like to hesitate but will grab what the market gives me , regardless of my initial PT.

    My PT and my SL have initial settings but are ALWAYS SUBJECT TO subsequent price action as it unfolds. Thus I trade with what I call dynamic SL’s and dynamic PT’s and they are based upon PA context over the last several bars, volatility, and subject to adjustment depending on the more immediate context of the present bar and last 2 or 3 bars.

    That is just the way I trade. I see the markets as very fluid and dynamic. I try to capitalize on the fluidity and the dynamics as they unfold. The market is full of uncertainty, and always will be, but that is part of the challenge. I am a strict daytrader so I am always flat by the close so I get my report card every day that I trade, by the close, at the latest.
     

  91. Thanks, its an interesting post and I'm going to have to go over it again more thoroughly.
     
  92. Can you post an example? I seem to remember attempting something like this a while ago and it wasn't helpful.
     
  93. When you see a thread in which nodoji and dxphoenix both apprise they trade this way,run for the hills.You won't find bigger misconception than when these two agreed to something.
     
  94. It is difficult to post market dynamics in a static post. Dynamics for me are watching price as it unfolds before and after I take a position, all within a context established by past recent price action. How do I illustrate that? I am not sure. I am an intraday trader that focuses on general context and every SINGLE bar dynamics as they are unfolding to make adjustments on the fly in terms of SL or PT. For instance, two bars in the ES that have a 2 or 3 point range. One forms slowly and closes in its lower half. The other makes it range rapidly with lots of back and forth price movement and closes on it’s high. As each one is forming the price dynamics are grinding in the first. The second one is formed with urgency and lots of price movement. I assess these two different dynamics with the context in which they are taking place and am constantly deciding if i should adjust my initial PT and increase it or decrease it and grab what the market gives me at the moment. If it isn’t clear to me i will generally opt to grab what the market gives (a bird in the hand is worth two in the bush) and if after my exit it appears I should have waited before exiting I will simply enter again. By grabbing my profits I avoid giving back a paper profit should the market prove my initial PT is not going to be reached. To me commissions are a small price to pay to grab a profit. Besides one is supporting the brokerage industry. They gotta make a living too...ROFL

    In a grinding uptrend (i.e. over several bars) what one can call a small pullback trend I will have multiple entries and exits. Instead of taking one entry and holding it. For instance, in a bull SPT I will exit at peaks and enter again at cheaper prices on PB’s. The SPT is the general context. On the pb’s I pay close attention to each bars dynamics to enter again. Likewise, on the peaks or swing highs for exiting.
     
  95. Not too knowledgeable about nodoji but dbphoenix is a charlatan. Ironfist is intelligent, though - if he says something I'd probably believe it. He is almost as critical of everyone as me, and that's a complete method in his thread with no guesswork or any other silliness that gurus here post.
     
  96. Can you post a chart?
     
  97. This thread with turn into a battle of those who believe that they can predict price action and those who believe it is impossible.

    Even those who feel as though they can read the market can have an exit signal countered by the following candlestick.

    So maybe trend followers should work on reading the market and price action readers should work on adding to a profitable position?

    Thanks to contributers in this thread as it has given me something to think about.
     
  98. I want to see an example of adding on every candle.
     

  99. Me too. Who does that?
     
  100. Me three. Who suggested that?
     
  101. Yea who said that? i.e. adding onto every candle? Schweiz talked about adding on with his grid...i think it was him??? Or was that grid discussion started by “salty” (morton). I just don’t remember.
     
  102. Are you asking if I know “how” to post a chart or would I post a chart illustrating the concepts?