Trading Systems in 'New Concepts In Technical Trading Systems' by J. Welles Wilder

[chirules54]

Hey Dale, interesting stuff about the linear regression above. After trying it out for a day or two, do you still think it has merit? I've been reading up on the RT system over the weekend, and I'm pretty sure I have it down pretty well now. One thing though, do you really think the "B, O, S" labeling really matters much? I know in our chat the other day when you talked about the "bastard version" that includes the S2 and L2 levels you never mentioned the day labeling so I wasn't sure if you still used that part of the system all the time. Since I don't really have much time to maintain and watch trades throughout the day, I'm probably going to takes trades based on where the candles close because for an hour or two after the "delta close" is the best time for me i.e. a break in between classes to look at the charts. So if a candle closes right above or below B1, I can set orders at LBOP, B1, and B2 with TP at MID or possibly S1.

Also, I hope to be trading live at Delta by thursday or friday. I withdrew the funds from my Oanda account and I am ready to sign up with Delta, I just need to find a bank statement to submit to them so they will approve my account.

[dpaterso]

Hello,

Geez: I was getting worried!!! No posts from anyone except me this week so far!!!

Well: I've added the LRC's to all my charts I definitely think they have merit and I've a few orders place based on this 'addition' so let's see what happens. In my 'research' into the use of the LRC's I found on more than one website that the combination of Pivot Points / Levels and LRC's is a 'valid' 'trading system' i.e. where the upper or the lower LRC line crosses a Pivot Point / Level then that's where to place your order(s) from what I understand. With these 'trading systems' you would TP at the LRC mid point. Now think about this for a minute: our RTS is Pivot Points / Levels as we have discovered so again we're not 'messing' with the 'core' of the system so nothing should change OTHER than the fact that you have a very clear indication of the existence of a trend or the absence thereof and how you 'play this' is of course up to you.

I DO think there is a CERTAIN merit in the 'BOS' sequencing and I can see the 'logic' behind it i.e. you buy or sell and then 'hold' for a day or two and this SHOULD have an impact on the amount of profit you make i.e. normally it would (should???) be greater. The success that I'm having with the RTS as you know is based on taking profit ANY DAY that the corresponding B1 or S1 is hit. Does this minimize the amount of profit that I'm taking??? Probably. Is it ensuring that I am INDEED taking profits and not losses??? I think so (even although the profit on a trade MAY be less had I waited for another day or two i.e. 'done nothing' on an 'O Day'). That's my 'take' anyway.

[balaji3003]

Dale,

Is it possible to make about 20-30% per month consistently by just trading the "main systems" in the book.....ie only trading at midnight when the new daily bar appears?

Thanks,
Balaji

[dpaterso]

Hi,

Yes I think so with the possible exception of the VS. The reason I say this is because the VS gets you into trades that may alternate between big profits and big losses so you may, for example, open a trade at the beginning of the month and that trade could very well sit in a loss for a quite a while until turning to profit and then once it's in profit you have the added problem that if you exit the trade early i.e. before waiting for a VS SAR then you cannot be sure just how long it's going to be before you get another VS entry signal. I do belive that a combination of the RTS and the SIS you should manage it quite easily if done correctly.

[dpaterso]

Hey,

Anyone out there who has a 'fundamental' knowledge of fundamental analysis???

Let me tell you why I ask:

I have just noticed a few things 'out of the ordinary' taking place so far today:

- The Dow is getting a good 'spanking' today i.e. right now it is down 176 points YET the GBP is gaining ground (the GBP normally follows the Dow ESPECIALLY GBP/JPY).
- Oil is going to yet another record high and pulling Gold with it YET AUD/??? pairs are not gaining proportionately.
- The USD is gaining ground in spite of the rise in Oil.
- The USD, GBP, and EUR are all gaining against the NOK (the NOK normally follows Oil).
- The USD is up against the TRY (the TRY normally follows Oil).
- The USD, GBP, and EUR are all gaining against the ZAR (the ZAR normally follows Gold).

As you all know I'm no expert in fundamental analysis (as a matter of fact I 'suck at it' BIG TIME) SO: are these 'divergences' indicative of anything or is what I'm seeing just 'wishful thinking' on my part ala AUD/NZD???

[balaji3003]

Can the reaction trend system be used to trade currencies?.
In the book, it states that "random price movement appears to take longer to increase than it does to decrease" - this applies to commodities right?

Currencies are traded against each other so i am unsure to whether the same principle applies for the currency market aswell.

If you have been using reaction trend system for currencies, how successful has it been(on average)?...win/lose ratio.

Also - how do you work out whether the first day is "B" or "S" day?.......i am asking this because im not sure how to work out the general trend(up trend or down trend)...ie im not sure whether to use the highest high or lowest low. - the ADX indicator fluctuates so much during a sideways market, it is not possible to see a clear trend direction.

How much money do you risk per trade? (in %)

Thank you

[dpaterso]

Good (Wednesday) morning all!!!

balaji3003:

IF your questions are directed at me then I'm afraid that you're NOT going to be pleased with my answer(s) to you.

I have gone to GREAT lengths on this thread to post details of what 'modifications' to the systems are needed SPECIFICALLY for forex pairs; what my % gains have been with the various systems; how to begin the 'BOS' sequence (ADX has nothing to do with this); and what my money management rules are i.e. 'how much money I risk per trade'. All you need to do is read the thread.

One thing that KEEPS coming up on ALL the threads:

WHY are people SO concerned with the win / lose ratio and / or the number pips made??? WHO CARES!!! We're NOT in this to make pips!!! We're NOT in this to attain the highest win / lose ratios!!! We're IN this to make $$$!!! People get 'stuck' on this 'number of pips' or 'win / lose ratio' for what reason I don't know!!! Remember: NOT ALL PIPS ARE CREATED EQUAL e.g. a 1 000 pip movement on EUR/USD is nowhere NEAR the $$$ value of a 1 000 pip movement on GBP/JPY or Soybeans or Silver!!! If I start the month with a capital balance of $500 and I end the month with a capital balance of $1 000 then I'm happy no matter whether the 100% gain on capital was made with a win / lose ratio of 1:50 or with a 10 pip nett gain!!! Remember: IT'S THE $$$ THAT COUNT!!! 'He' who has the most $$$ WINS NOT 'He' who has the most pips NOR 'He' that has the highest win / lose ratio!!!

[chirules54]

Dale, I've been spending more time looking at linear regression channels, and I do think that it would be a better "compass" for the RT system trades. My main problem with the RT as it is written in the book is which day is designated as your previous significant high/low point. Should you start looking when the ranging period starts (adx and adxr both under 25) or at some previous point? Also, I have noticed that the B O S sequence can sometimes get messed up when you do it like it says in the book. I think this is because the system is based on the "three days up, two days down" thing mentioned on page 72 and 73, which doesn't seem to happen as much nowdays, just from my looking at some charts of low adx/adxr pairs recently.

Also, I was wondering what settings for linear regression you were using for your RT entries. To me, what seems to work pretty good is to set the linear regression lines to start at where the range seems to start (first day in a sequence where adx/adxr values are below 25) and then possibly extend the lines by adding 1 to the value every day.

Not trying to get away from "the man's" systems or anything, but if you use linear regression by itself like I mentioned above, you seem to get pretty good results selling off the top line, buying off the bottom, and taking profits at the midline. The trick seems to be setting the linear regression to start where the range starts, and then adjust it every day by adding 1 until the adx/adxr values start to climb back above 25 and you no longer want to use an anti-trend system.

[dpaterso]

Yo!!!

'The book' says to go back two or three weeks and SOMEWHERE AROUND THERE find a significant high point or significant low point and start the sequence from there and it does not matter whether the instrument of pair was ranging or trending BACK THEN ONLY what it's doing now. Also remember that the 'three-days-up-two-days-down' 'phenomemon' I don't think is supposed to be taken 'literally' as it were. In other words: remember that you're looking back HISTORICALLY i.e. you're seeing the bars AFTER they've closed. You're (we're) trying to trade a ranging market so EVEN IF you have FIVE 'down days' in a row it does NOT mean to say that DURING those five 'down days' the RTS levels were not hit. Remember: they're Pivots and 99.999% of the time price will trade to, and stop, and turn again at one or the other Pivot i.e. at one or the other RTS level.

As far as the LRC's are concerned: I'm 'with you' on your idea of finding 'roughly' where the instrument / pair started trading in a range. What I'm currently testing with LRC's (14,100) is to find a pair that's trading in a range and placing my buy or sell orders at the FIRST OR CLOSEST RTS level that falls OUTSIDE of the LRC's and then following the RTS through for an exit. In other words: if my B2 is the closest RTS level AND it's OUTSIDE the LRC i.e. BELOW the red line in our case then that's where I'm placing an order and if it gets hit then I'm making my TP S1 as per the RTS and adjusting that every day if the TP has not yet been hit. Yes: I'm making fewer trades of course BUT I'm ALMOST ensuring that if my TP i.e. S1 in this case is NOT hit on the 'correct RTS day' and price moves against me then the next day, even although S1 has moved CLOSER to me, if I manage to TP then I still make a profit as opposed to either breaking even or having to close out at a loss when I'm required to close out as per the system. I hope that makes sense.

You guys seem to be having the same problem as I had with the book (and no offense is meant here i.e. this is meant to help not 'hinder'): I don't know if it's because the information is presented in such a 'dense' or 'concise' form or what but sometimes you need to read each sentence as if you were back in 'kindergarten' because it's REAL easy to miss a word or something that can change the whole outcome of your understanding of what you've just read.

Also: I am currently 'experimenting' with using ADX/ADXR(7) again to detect a range. The reason being that these RTS trades appear to be relatively 'quick' trades so you want to know ASAP when a pair has started trading in a range and you also want to know ASAP when a pair has started to trend. I'm currently not sure whether I'm on the right track or not i.e. only time and experimentation (and resultant profits of course) will tell.

Again: every day I find more and more that you need to 'see' the systems 'as a whole' i.e. take a 'broader view' of things e.g. don't JUST look at the ASI or the RTS levels BUT ALSO take into account the 'slope' and 'direction' of the ADX and where +DI and -DI crossed because that is also a 'key' level (take any instrument or pair, see where +DI and -DI crossed, and look at the closing price on that day keeping in mind the 'extreme point rule' and you'll see that most times that level is either support or resistance). Just one example of course but you get the picture.

Edit:

I thought I'd attach a chart of USD/ZAR (ADX/ADXR < 20) 'for good measure'.

[dpaterso]

Hey Nick,

Seeing that you and I are 'playing around' with LRC's here is what APPEARS to be the MOST 'foolproof of the lot' and I thought you might be interested or at least take a look:

LRC's (14,100).
ADX/ADXR > 25 (the higher the better of course).
RTS.
SIS.

Take SIS entries (whether based on HSP's or LSP's or SAR's it does not matter) ONLY in the direction of the LRC's AND (here is the 'key') EXIT at the closest RTS level that will result in a profit being made (or if you're feeling 'adventurous' scale out the position over multpile RTS levels!!!

In other words: you're using the SIS for entries (and SIS entries are 99.999% 'accurately in the right direction' INITIALLY I find), you're ONLY taking entries in the direction of the current trend, and you're taking profit at specified levels as per the RTS. Given the numerous entry signals provided on a daily basis by the SIS this should generate nice consistent daily profits EVEN ALTHOUGH you're not using the SIS to 'trade the trend out'.

This of course is NOT 'trend following' and works on a similar principle to the TBPS except that the profits are bigger, you're NOT entering at market, and there are no HUGE stops to spoil you day (account)!!! On the OTHER HAND: the LRC's sure look a like a MAJOR 'compliment' to the TBPS i.e. only taking TBPS entries when indicated BUT in the direction of the LRC's.

I suppose that all of the above once again just proves the old adage that 'the trend is your friend' (as 'coy' as that may sound after all this time)!!!

Take a look.

Edit:

On FURTHER closer inspection it would seem that by decreasing the period for the LRC's they become more 'sensitive' and / or 'reactive' to a trend direction change. Could it be a way to 'filter out' some 'false' SIS SAR's???

(Just one thing for others following this thread: Linear Regression Channels or 'LRC's are NOT just a set of parallel trend lines i.e. their values are calculated by using a (fairly???) comples algebraic equation. I mention this because I think that they have more 'substance' than just a set of parallel trend lines).