You are very welcome :)
Printable View
You are very welcome :)
Thks for the update Trackers..
I'm curious though...... why you are testing the Stochs ..There are many other much better indicators around these days..and the MSI in my opinion is one of the best, with a far better record of success (risk V reward)...
I'm biased when it comes to Stochs so my comments maybe unjust..but as I see them the stochs react to price with momentum so why not just look at the price movements on big trading days....Stochs look great in hindsight
when they get very high (90+) and the cross over occurs resulting in divergence warning of a new future price fall...but that not the case when the Stochs are low (10-) as there is more price volatility resulting in unpredictable
Whipsaw activity...An example on your above chart would be buying in during a lull with the Stoch signaling a divergence (a bottom?) but in reality it is still a continuing downtrend (22 Sept and 6 Oct periods)...
I see you have adjusted the line to a more sensitive setting of 30 to enter quicker after a bottom...
As always a quicker entry involves a higher risk....therefore there will be difference with the MSI medium term settings
MSI incorporates many indicators and when Phadreus (P) was in the process of creating it he mentioned a few tweaked things to me (via PM) which I agreed with a couple of words so no real conversation took place...but in those
early days he gave a me a formula to look at. It had adjusted indicators with 10 results, two of which incorporated the Stochastics...He tweaked it a lot since then so I dunno the final makeup of the MSI. I assume the makeup is
using medium term settings, as the MSI is nearly spot on when comparing with other isolated indicators or chart patterns...and so it should be, as P designed MSI to be the ultimate indicator and such has the unique ability to
be used in isolation..Try charting..drawing trend lines, S&R lines on that colour coded chart, try adding other medium term TA indicators..as you will see all those differing disciplines will all agree with the MSI... so obviously the
colour coded MSI price bars (or lines) creates a simple chart thereby eliminating the need to clutter up a chart with lines and boxed indicators.
So Trackers that's my argument... there's no need to include anything else...eh?..If you want to predict an entry closer to the bottom (more sensitive), why not just create a shorter term MSI chart with shorter "n" periods...
and forget about the Stochs..
Note though the shorter the time frame the more active (shorter term discipline) your trading has to become in response to more signals so as not to increase your risk ..Using quick signal indicators with a medium term
investor discipline upsets that discipline system and increases risk..as well de-optimising that discipline's risk v reward ratio
Your thoughts???
Hoop
EDIT: Just saw KW's post...agree MSI keeps you out of bearish contining event patterns such as the dead cat bounce...also I remember P desensitised the MSI enough so not to "chuck you out" with moderate bull market cycle corrections
and signal strong buys during moderate Bear market cycle corrections (sucker rallies)
Hi guys,
Well a couple of things. Firstly, Phaedrus never gave anyone his formula. He did however point out the components of his formula on a few occasions, and I spent a hell of a long time working out how to code Metastock to pull them together into a points system and one common indicator - Luckily, I came up with it long enough before he stopped posting that I could see that it was identical as time went on, and he confirmed a couple things with the Expert Advisor to me in a PM, and I also told him to let me know if he had an issue with posting it in his absence.
Hoop, you're right, when it first was created it was something like 5 indicators, with two variables per indicator and an overall average, but ended up being the average of 50 calculations (and maybe slightly higher but with no discernable difference) - You could make it more responsive by leaving out the longest 5 or 10 calculations I guess
He also confirmed to me in a PMs that backtesting for profitability was ultimately *the* way to come up with an indicator, and of course that makes sense. You do this by backtesting a bunch of standard and custom indicators (using variable time periods, and variable parameters) over the market (or individual stocks, or whatever), over a large enough period of time, and see which are the most profitable.
As he did this, obviously a couple rose to the top (they did for me too), and then he made some custom indicators which cobbled a few of these together and backtested again, and the custom indicator outperformed the indicators in isolation. Easy
The thing is though, the last time I talked to him was almost 4.5 years ago. If he backtested the MSI now, versus all the other common indicators available, with the benefit of having an extra 4 years data... Would the MSI still outperform everything else? Well, it is a great indicator that is for sure that is extremely safe and really stood the test of time, but the answer is definitely not certain, and that's why other indicators should always be considered (at least in combination). You also have to consider a number of other things when thinking about an ultimate indicator (and Hoop a lot of this is what you're saying and I agree):
1. The most profitable may involve far too many trades, and even though from a pure monetary sense it results in the most $$ at the end (system includes customisable trade fees), if there's a system that's close but has a fraction of the trades, its probably better
2. Winning versus losing trades. If the most profitable system has only a few winners, but those winners are massive, if there is another system that is close but has a better win %, that is probably better
3. What stocks are you backtesting against? Small caps? Everything?
4. How far are you backtesting. what if you backtested against different time frames? Would the best performing indicator still outperform
etc
Recently I backtested everything (MSI, stoch, all the usual suspects) against 50 stocks in the ASX 200 going back a year. Stoch significantly outperformed everything else. That's a very short time frame, and a very small number of stocks, but as a quick bang up in 2 seconds indicator that'll *most likely* produce a reasonable result in the current environment, it seems pretty good (and Stoch is a component of MSI as Hoop correctly says) - A lot more work needs done though
Trackers...Your post cleared up some misconceptions I had...I assumed wrongly when I asked P what was going to happen to his MSI after he stopped posting on ST..In my view it was a too good of indicator to be forgotten about in the TA world
P said that you had metastock and he was showing you the formulae to instal it as you showed keenness to carry it on...I knew P was still continously tweeking his MSI..I took that to indicate that he was passing his version onto you......
So sorry if I called your MSI as P's MSI.....maybe we all should call it MSI Tracker or something like that..to show appreciation to the TA world that this is your individual version...In my opinion, the amount of time one spends on creating
a reliable indicator they should be recognised for it...
So thanks again Trackers, and I look forward to see your outcome on those Stochs..
I too would like to thank Trackers,and you too Hoop for the hard work you put in ,and the fact that you both share your research with us.
Trackers I trust you are on the mend.
Hey Hoop.
So I backtested the ASX200 going back to 2008, and the Stoch beat out MSI, as follows:
MSI, average profits between $20,242 and $20,475 (depending on parameters)
Stoch, average profits between $26,563 and $27,227 (depending on parameters)
It was actually worse than that, but I took out SIR, SFR and a couple others that were skewing the results big time in favor of Stoch
Stoch benefits from optimisation more but even completely picking random parameters it would be (much) more profitable. A funny thing with MSI is the system optimised the buy parameters all over the show with little difference in overall profit, but was convinced the best sell parameter was -1.
The Stoch optimisation was weird as well, putting the best sell indicator in at a cross of around 5 or 10, which was a bit bizarre.
So I made a system that bought using Stoch, and sold using MSI, and that did really well, average profits between $28,978 and $29,980.
This is all using the benefit of hindsight, but in this case hindsight would only tell you what indicator to use; how you used it (within reason) would then not affect the result in any way
And thanks Percy and Sharer, appreciate the kind words :) And Percy I can walk again, yay! :D
Great news.