Strategy development

[barjon]

Did you define your criteria/setup entry somewhere? I must have missed it.

dpb

No I didn't. For the example trade I was using sulong's breakout/pullback setup together with some further rationale which I posted at the time.

jon

 

[dbphoenix]

dpb

No I didn't. For the example trade I was using sulong's breakout/pullback setup together with some further rationale which I posted at the time.

jon

That creates problems with regard to trade mangement since the manner in which one manages the trade will depend on the testing he has done of the setup.

 

[barjon]

dpb

in haste - just dashing out.

I have done a lot of testing - although I can find (and have) one or two objective rules, I find that much relies on judgement of the price action which is not easy to be objective about unless one has considerable experience of the price/volume dynamic. Does that make sense?

jon

 

[dbphoenix]

I have done a lot of testing - although I can find (and have) one or two objective rules, I find that much relies on judgement of the price action which is not easy to be objective about unless one has considerable experience of the price/volume dynamic. Does that make sense?

My comment had to do with the question of whether or not you had tested sulong's setup, since that is the one you used. What is it that you've done a lot of testing of and what variables did you test? (if you still can't be objective about the trade, then perhaps there were important criteria that your testing did not address)

 

[barjon]

dpb

still here - the bridge table calls in 16minutes 35secs!!

No, I hadn't tested sulong's set up - just making a contribution to the thread. My main trading is swing trading - position, not intraday. I could detail the variables, but I don't want to litter the thread since it's off topic. Can pm you if you're interested.

jon

 

[dbphoenix]

dpb
No, I hadn't tested sulong's set up - just making a contribution to the thread. My main trading is swing trading - position, not intraday. I could detail the variables, but I don't want to litter the thread since it's off topic.

I don't think the bar interval and the setup are related, i.e., the same setup occurs on daily charts, weekly charts, etc. It's up to sulong, tho.

Of course, if you're not interested in the setup, then I'm lost again.

 

[barjon]

I don't think the bar interval and the setup are related, i.e., the same setup occurs on daily charts, weekly charts, etc. It's up to sulong, tho.

Of course, if you're not interested in the setup, then I'm lost again.

dpb

back again.

I'm not quite with you . Agree that the bar intervals don't matter - I was just explaining my main type of trading.

I am interested in the breakout/pullback set up which is the main subject of this thread, but it is not one I have tested to any significant degree. I have tested, extensively, the setups I use for my swing trading - I thought that was what you were after when you asked "....what have you done a lot of testing of...".

Just to be sure we're not talking at cross purposes I am defining (there's that word again ) setups as taking us as far as entry with managing the trade/exit strategy as a subsequent and separate feature.

regards

jon

 

[dbphoenix]

Just to be sure we're not talking at cross purposes I am defining (there's that word again ) setups as taking us as far as entry with managing the trade/exit strategy as a subsequent and separate feature.

If trade management is a separate issue, then what is it that you're testing when you test a setup (i.e., the setups you've tested extensively)?

 

[barjon]

dpb

I test the setups to find out if, and how far, the price moves in the right direction after entry and over what timescale. This only demonstrates the potential profit to be gained - trade management determines how much of that potential you actually realise.

This is becoming something of a dialogue that I doubt interests many readers - do you want to continue via pm?

regards

jon

 

[dbphoenix]

dpb

I test the setups to find out if, and how far, the price moves in the right direction after entry and over what timescale. This only demonstrates the potential profit to be gained - trade management determines how much of that potential you actually realise.

This is becoming something of a dialogue that I doubt interests many readers - do you want to continue via pm?

regards

jon

I don't see any point in discussing it in private. The fact that traders are so reluctant to talk about definition and testing generates an awful lot of unnecessary posts which skirt the subject without ever really saying anything about it, similar to the way people used to talk about sex. "Moving on" to trade management just shoves the original subject back under the bed.

But, again, it's sulong's thread, so I'll put all of this back where I found it and wish everybody well.

 

[sulong]

dpb

I test the setups to find out if, and how far, the price moves in the right direction after entry and over what timescale. This only demonstrates the potential profit to be gained - trade management determines how much of that potential you actually realise.

This is becoming something of a dialogue that I doubt interests many readers - do you want to continue via pm?

regards

jon

Jon,
I think the subject of testing is very much appropriate to strategy development, including "trade management".
I've had /have difficulty's in this area, and so I'm always looking at that part of my plan, trying to discover where my weak area's are.
I'm sure that this type of dialogue is of interest to all those that participate in testing out their strategies.
For what it's worth.......

 

[Quercus]

The way I see it, once I have a DEFINED set-up, is to look at the average profit and use that as a target. There will be other ways to manage the trade, but for me, with my inability to hold trades, I need some evidence that the probabilities say that my target should be "x". I believe this should help overcome some of the "psyche" issues.
I am still looking and getting sore eyes!
Cheers
Q

 

[JumpOff]

..The fact that traders are so reluctant to talk about definition and testing generates an awful lot of unnecessary posts which skirt the subject without ever really saying anything about it, ....

I suspect there are several reasons for this reluctance.

One of them is that many traders have gained a "feel" for what their favorite setup looks like, and how it reacts, but find it difficult to quantify mathematically.

Another is the fear of putting hard won knowledge into the public domain. If I was able to mathematically specify a setup, stop and target that brought me pleasant gains and little draw downs, would I blather it on these boards?

Of course there is the possibility that folks won't tell, because they really haven't got a clue, and just 'think' they have a plan...

And let's not forget the lurkers and wannabe's ( I put my self in this category) who won't post because they know they don't know.

JO

 

[barjon]

Jon,
I think the subject of testing is very much appropriate to strategy development, including "trade management".
I've had /have difficulty's in this area, and so I'm always looking at that part of my plan, trying to discover where my weak area's are.
I'm sure that this type of dialogue is of interest to all those that participate in testing out their strategies.
For what it's worth.......

sulong

ok, but I hope it doesn't put people off any further discussion of entry setups.

The key elements seem to me: 1. Initial stoploss; 2. Adding to position(s); 3. Protecting profit stops; 4. Final exit(s). Maybe talk about each in turn to avoid massive posts?

Initial Stoploss alternatives:

1. The lowest risk in terms of money on the table is to exit immediately if the price moves against you and to put a stop at breakeven as soon as that's possible if it doesn't.

2. Put a stoploss just (define just!!) below the low of the last pullback bar.

3. Put a stoploss just below the breakout point ie: the previous major resistance that should now become support.

4. Have no stoploss, but rely on your analysis of the price/volume action to determine if you should exit at a loss.

If I had the skill I would probably like to use 4, but I don't so I am more mechanical. I use 1,2 or 3 (as far as applicable to swing trading rather than the breakout/pullback topic of this thread) depending on the sort of price action (volatility) that seems "usual" for the stock in question and depending on the level of risk (eg: below the breakout point may be just too far away). A problem with 2 and 3 is that stops are obviously placed and often gunned - the conspiracy adherents would say by design, others would say no conspiracy, you just put them in the wrong place

A starter for ten anyway.

Good trading

jon

 

[tune]

I have been reading this thread with great interest because learning how to develop and test strategies is precisely what I want to do.

Below is a set-up that I am trying to hammer out that the moment. It is similar to the one at the beginning of this thread. Make what ever comments you like about what I have written. I have thick skin and I am here to learn.

IF an extended base has formed, after a downtrend, at previous support
THEN price breaks up above the R of this base on strong volume
AND there is a low volume pullback
THEN place a buy order just above that pullback.
(Place a sell stop at the same time as you place the buy order, put it underneath the base support level.)
IF the price continues to rise then move stops up.
(Put them underneath each higher low so that if price falls below the last higher low you will be stopped out of the trade)
Continue to do this until you are stopped out of the trade.

The reason for putting the buy stop above the pullback is to get stopped into the move only if the down bars do become the type of pullback that I am looking for. I don't want to buy a reversal. I am looking for the same type of pullback that I have seen hundreds of times. The one that goes wide bars up, narrow bars down, expanding volume on the up bars, contracting volume on the down bars. Regular stairstep.

Obviously this is not an accurate IF THEN. It's only the bones. I presume that testing is required to define the details. I do not know how to do this as yet. Perhaps my set-up is still too loose to be tested?

If my pullback does not become a continuation and instead becomes a reversal my trade has failed. A fade IF THEN plan is something I am thinking about.

The type of set-up I am looking for is, to me, a continuation. But to some people the little down move I am aiming to hover above may be something else. In ticks my pullback could be a reversal, the whole downtrend, base and breakout into an uptrend could also be a reversal. However, looking at a bigger picture this reversal could just be one part of a low volume pullback after an upbreak from an extended base. Is this what is meant by forests and trees?

tune

 

[dbphoenix]

The type of set-up I am looking for is, to me, a continuation. But to some people the little down move I am aiming to hover above may be something else. In ticks my pullback could be a reversal, the whole downtrend, base and breakout into an uptrend could also be a reversal. However, looking at a bigger picture this reversal could just be one part of a low volume pullback after an upbreak from an extended base. Is this what is meant by forests and tress?

tune

I wasn't going to re-enter this particular fray, but since you've made the effort to answer sulong's original question, congratulations are in order. All other considerations flow from the ability (and the willingness) to distinguish between a retracement and a reversal. If one hasn't done that, then he's going to be swimming with ankle weights. In a heavy robe. Tied to the pier.

You've made an excellent start by defining the context, and I'm assuming that if this particular setup were not to occur off a base at potential S (note I said "potential") after a downtrend, then you would not consider it valid.

However, you're getting tangled up in the politically-correct definitions of reversal, retracement, pullback, continuation, trend, breakout, etc. I suggest that you avoid this since these are all only labels used to describe what it is you're seeing in the charts. What matters is what you see, not what you call it, at least until you want to discuss it with somebody else, in which case you simply provide your own set of definitions. In other words, you decide the rules of play. If someone wants to talk to you, they learn your dialect.

So.

You say "the type of set-up I am looking for is, to me, a continuation". Note the words "to me". And if this is how you want to define "continuation" in your own mind in order to sort all this out and make sense of the tactic set you're trying to develop, then that's fine. But first you need to think about what it is that you want to see continue. I assume that you mean the direction in which you want to see price move, in which case, you're going to have to think about the context.

For example, if you consider the up and down movements within a trading range to be uptrends and downtrends, then the bounces off S and R might be considered reversals. If the subsequent move is a trend, then any counter-trend movement is a retracement, and any resumption of the trend is a continuation. However, if the movements up and down within a trading range are not trends but just bounces back and forth, then any entry you make is part of the previous reversal and need not have anything to do with retracements and continuations, depending on how you're defining all these terms.

What difference does all of this make? Probably none, if you're a scalper. But if you want more than that, you're going to have to consider probabilities, factor in your price targets, the width of your stops, your win:lose and profit:loss ratios, the strength of your initial setup, etc. For instance, if you can enter the trade very near S, and R is enough of a distance away to make the whole thing worthwhile, that will be more attractive than an entry made halfway through the trip from S to R, much less an entry made very near to R, i.e., your target (perhaps).

On the other hand, if you decide that trends take place only outside of ranges and that retracements take place only after price has broken out of the range, a new set of possibilities is available to you, not the least of which is that you have no set price target. You may decide that that "retracement" that you're buying after the bounce off S is just a higher low and of no particularly cosmic significance. If price wends its way toward R, fine. If not, you've planned for that contingency. But by not thinking of the move as a "trend", you avoid some of the usual pitfalls, such as confusing trend with chop, and getting lost in increasingly smaller up and down movements.

There's more, of course. But this is something to chew on over the weekend.

 

[Quercus]

Great post dbp!
Thanks
Q

 

[tune]

Thank you for your comments Db. I did not understand every word but will persist until I do. Your point about providing my own set of definitions is not something I had thought of and I am more than happy to learn other dialects in order to talk, if only because I am always on the look out for more, apt trading words. I have recently started watching live charts and as much as I enjoy the movie, the narrator is not up to much.

Your assumption about my set-up is correct. I am looking into other range based set-ups but the one I have tried to describe is the only one I am referring to. And yes, I do note that you have said "potential" S. I have been calling it "possible" S to myself.

Part of the reason that I am getting tangled up with politically-correct definitions is my desire to make complete sense of post (#177). I know that there is something vital here but I do not fully understand what it is as yet.

For a trader, a definition enables him to make a choice. Defining a reversal as a change in direction, for example, or as a break of a TL or whatever, is useless. It says nothing about what the trader is supposed to do. A trader has to distinguish between retracements and reversals in a way that tells him what to do. If he hasn't done that, then he's wasted his time.

The following is my thinking on it so far -

Going back to your example of considering the ups and downs within a range; my preference would be to think of these up and down movements only as individual uptrends and downtrends if the range was of a "decent" tradable height (something like at least 15 points on the NQ). i.e. enough elbow room for visible stair step trends to develop. If the range was only 5 or so NQ points, that's just trend less rattle. It's way too tight for me and I would not attempt to buy the bottom of it in order to try and sell at the top. Or vice versa.

Calling my set-up a continuation move makes no sense if the range is only a few points high. Until there are higher lows and higher highs above this range (congestion might be a better word) it would not be accurate to start thinking in terms of trend and counter trend. The "pullback" I have been looking out for is, so far, only one higher low above some congestion. One higher low does not constitute a trend. If a higher high develops after this low, then it can be considered as a trend (I think).

I see what you mean by entering close to S rather than above R . An above R trade would be more expensive in terms of stops (assuming you still wanted to put the stop under S) and it will have already travelled some way up. The trade down near S might cost less than the one just above R, but this does not mean that it will necessarily be better value because the probability of the above R trade succeeding could be much higher. The near S trade though, if the range was narrow, would just buy you straight into the congestion, which was not the idea in the first place. With a wide trading range, however, it is quite different, because in this context R is now a good few points above and it may be worth the risk.

tune

 

[sulong]

Part of the reason that I am getting tangled up with politically-correct definitions is my desire to make complete sense of post (#177). I know that there is something vital here but I do not fully understand what it is as yet.

For a trader, a definition enables him to make a choice. Defining a reversal as a change in direction, for example, or as a break of a TL or whatever, is useless. It says nothing about what the trader is supposed to do. A trader has to distinguish between retracements and reversals in a way that tells him what to do. If he hasn't done that, then he's wasted his time.
tune

I don't know if this will confuse things even more, or maybe help, but here it is.

My current project is putting the different pb's/ret's setup's in classes.
The idea comes from Alexander Elders book "trading for a living". where he made up 3 classes of divergences.
My idea is to have a cheat sheet, that will tell me what the highest,second highest, third highest probable pb's/ret's look like coming from
A. BO from horizontal consolidation.
B BO from a consolidation of lower high's and higher lows. ( triangle's)
C. Reversals. ( yes, for me I see pb's from reversals.)

The end result I have in mind, is to allow myself to only take the very best trades out of the many possibilities that I see. (3 setups out of 9)

Although, it's turning out to be a pretty large project, we'll have to see how it ends up.

 

[tune]

Thank you for your reply Sulong. I don't know whether it confuses my issue or not. I'm not even sure I have an issue. Perhaps I already do take responsibility for the trade but am not aware of doing it. I'll figure it out in the end.

It sounds like an interesting project you have taken on. Anything which helps work out which set-ups put in the best performance has got to be a good thing.

Are you deducting percentages manually or by backtesting with a computer? I presume you are categorising types by looking into things like, whether the retracement low/high fouls the previous low/high or not, and by how much?

tune