what performs best?
Author: dannyg
Creation Date: 6/25/2008 8:37 AM
profile picture

dannyg

#1
just downloaded wealth lab and looked at the installed strategies which all seem to perform worse than a buy and hold position when using the index data field....am I missing something?
which strategy actually works better (and doesn't involve 100's of trades)?

looking for something that I could correlate with option credit spreads on a monthly basis.
profile picture

swuzy

#2
I don't know about options, but you need to run the WatchList Rankings comparisons on your watch list. You should install WL on a spare computer though, one with lots of ram; because it will tie up your computer for several days to rank all available public scripts against your watch list.

Alternatively, or simultanuously, have you noticed that the archived website has a Search http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=ChartScriptSearch.htm window under ChartScript Center in the left hand column. In that Search window, scroll down and you can input various parameters to search for .... for example Annual Profit Rate in excess of x%, Drawdowns not more than y%. These Search rankings are on the standardized WL 100 watch list.

profile picture

dannyg

#3
what I was looking for is a comparison of the installed strategies showing which actually work better than buy and hold...are there any? I know time frame has something to do with it but is there a place that does some comparison?
profile picture

swuzy

#4
Recent edits are in italics:

On the archived website, http://wl4.wealth-lab.com/, just scroll down to the bottom of the page for a quick top ten list by net profit.

Tables of top 25 rankings by net profit are here http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=Top25APR.htm

Tables of top 25 rankings by Wealth Lab score, of high scoring, but "safer" and having relatively low number of trades are here http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=Top25.htm.


Table of 20 most often ran Chartscripts on the website in the last 7 days and labled "Most Popular" are listed here http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=Popular.htm

Problem is it all varies depending on your selected set of watch list. Your preferences in stocks going into a watch list will have a commonality of optimal entry and exit points that will be different from another set of watch lists.

The website's default is to run search criteria of desired features of scripts against the standardized Wealth Lab 100. This is a very difficult set of symbols to do very well on. Yet, those scripts that do well on the WL 100 might not do as well on the NASD 100, or do worse on the IBD 100; and those scripts that do well on the IBD 100 may not do well on the WL 100 or the SP 500. So scripts that do very well on your list may not do well against any of the other standardized lists.

You can also start by running scripts high in popularity ranking against your favored lists.

Yes, scripts that do well on my favored list, generally do not do well w/ the WL100, but I continue to work on trying to improve them all around.

Using the site's Search Rankings http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=ChartScriptSearch.htm feature is a start.

The best way is to run the ChartScript Rank feature (in the stand alone WL application) to rank all available chart scripts against your preferred lists. Then make some choices, analyze the scripts you like, and try to improve on them.

As I said for the ChartScript Ranking program, you will need a PC with a lot of ram, I know 1 GB resulted in locking up my old PC after a partial run. Then I ran it on a 4 GB PC running a reasonably fast dual processor and it took about 3 to 4 days and night before finishing the batch.

Also, most or almost all of the scripts are dip buyers. So if you run them on one or a handful of strongly trending stocks, they won't do very well. So if you favor strongly trending stocks, you need to run the scripts against a large list of strongly trending stocks, and then the scripts will have greater opportunities to cycle in and cycle out and generate better returns. On the other hand, such scripts will likely generate a very large number of trades requiring constant attention.

Scripts that do not generate a large number of trades when tested and ranked on WL 100, will have a lower annual return, but also a very low Exposure (time in market) and might work better for you, but you have to test it against your own lists, and you may want to try to adjust and select various position sizing on such types of scripts. For such scripts, you might run several scripts together (applying different entry and exit strategies) in the MultiScript window on your watch list to apply a more full spectrum press.

Many listed scripts focus on a single or relatively fewer entry exit strateg(ies) to facilitate testing, ease of understanding. This also helps one make modularized choices to allow for selecting several to run together on a list. This should be considered a feature.

After a while, you can write your own script combining several of these strategies into a single favored script.
profile picture

swuzy

#5
Note that some of the chartscripts with "the highest returns," are actually invalid scripts. Many such are concept scripts which optimize themselves, some by means that "peek into the future." When you run a two or five year back test, they will optimize their indicators' parameters by looking into the two or five year period future to the tested date. So you need to read their commentary or their attendant icons to avoid them.

Note also that if you have a highly selected set of symbols, you may generate a very good back test, but that script you are using may not be able to produce results any where near what you tested. This is because your selected list of symbols has a "survivorship bias." For example, if you do a two or five year back test on your selected symbols or say the IBD 100, you start on day one, that is two or five years ago. But your symbol list already has tossed out all the bad stocks for the next two or five years, and kept only the best stocks that survived your criterias over the next two or five years. So you have built into your back test, a systematic peek into the future, guaranteeing you a wonderful backtest result. However your script may not then be able to sustain its performance in the future, because your stocks may not be able to sustain their high performance into the future.

This is why standardized testing use the Wealth Lab 100, which is a more randomized set of symbols, and that is why it is much more difficult to achieve good back test results when running scripts against the WL 100 as compared to an optimized list such as the IBD 100.

profile picture

swuzy

#6
Being a nonprogrammer and a relative neophyte, I don't pretend to know what might run best under different lists and different circumstances.

However, after having shared the various thoughts above, I can't resist to offer a BUT THEN, TRY THIS, perspective with some TRULY EYE POPPING RESULTS.

In WLP 4.5, set up your $imulator for $100,000 starting capital, 10% equity positioning, 20 year period. First run the 33 Liner chartscript on the IBD 100; Then run the same 33 Liner chartscript on the WealthLab 100.

You will be quite astounded at the comparative results; especially for the WL 100. As of this date's run, the IBD 100 ends up with about $5.3 million of ending capital (no taxes taken out); But the WL 100 ends up with over $1.1 billion of ending capital (no taxes taken out), an annualized gain of over 100%.

On my personal watch list (which I consider to be somewhat bloated from lax of discipline and needs trimming), the same parameters leave me with ending capital of over $51 billion (annualized gain of 186%, no taxes taken out). Such wishful thinking.

So why am I losing my shirt in this stupid market?

Notice the 68% maximum drawdown on the WL 100; never play long a down market; only suckers and amateurs like me think against all evidence to the contrary, that they can get or might have an edge going long in a down market. Obviously, it is not the market which is stupid, it is me who is stupid.

Cheers, good luck, and don't lose the money you take off my back.


profile picture

swuzy

#7
So, ..... can some one please please provide a copy of the 33 Liner translated into WL 5's code? Thanks so very much.
profile picture

Eugene

#8
Here it goes. Keep in mind that it does not follow the original script in terms of exits. It's just my view how they should work.

CODE:
Please log in to see this code.
profile picture

swuzy

#9
Thank you, thank you Eugene. You are sooo goood.
profile picture

gharakhanian

#10
Hello SWUSY,
I am new 5.0 .net user and have just started playing with it. Could you please help me figure out the answer to the following questions? I inserted your comments from above first followed by my question:

You Wrote:
Tables of top 25 rankings by net profit are here http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=Top25APR.htm

Question:
Is there a way to download the .net translation of the above 25 to my environment?


You Wrote:
The best way is to run the ChartScript Rank feature (in the stand alone WL application) to rank all available chart scripts against your preferred lists. Then make some choices, analyze the scripts you like, and try to improve on them.

Question:
How do you invoke the "chartscript rank" feature?

Another Question:
Is there a way to build a macro which runs a script many times with random start and stop dates and averages the total returns.

Al

profile picture

Eugene

#11
QUOTE:
Is there a way to download the .net translation of the above 25 to my environment?


Ask Fidelity about their chartscript conversion utility. It was supposed to do this automatically (when released). The only other way is manual translation.

QUOTE:
How do you invoke the "chartscript rank" feature?


It was a WL4 tool yet to make its way into WL5 in a later build (5.2).
profile picture

Bandit292

#12
Eugene - ref the 33 liner test code. I see the WL100 as an existing data set. How do I import the IBD100 as a data set, or do I have to create it and enter each stock symbol manually?
profile picture

swuzy

#13
Al, you asked,
QUOTE:
Is there a way to build a macro which runs a script many times with random start and stop dates and averages the total returns?


I am certain that can be done, but it is beyond me for now. If you raise it as a separate request with a higher degree of specs, someone might be able to help you.

WL 4 or WL 5?

Because of various aspects of WL5, I think "beginners and non programmers" probably should still stick to WL 4.5 for now for the heavy lifting. You will have a lot of public scripts to play with, and it is a mature product, well known to many aficionados. If you have not installed WLP 4.5, you should definitely do so and rely on it as your primary engine. WL 5, at this moment, cannot convert and cannot run any of the many hundreds if not thousands of public Chartscripts created prior to WL 5.

IBD 100.

I suspect you cannot automatically import IBD 100 as a data set. You can get IBD 100 updates every Friday late afternoon directly from the IBD investors.com subscription website here http://www.investors.com/member/ibd100/viewibd100.asp (it loads into Excel and you can Control C copy the 100 symbols from there), or wait for it to be updated by rhino or others in the public watchlist site to which you must log in, and then Select and Control C copy it from there http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=PublicWatchLists.htm

It is not difficult to create an IBD 100 data set for WL 4.5 or 5 and update the symbol set manually weekly. I would create it and update it in WL 4.5 and in WL 5 create an IBD data set that is sourced in WL 4 so that I do not have to update it twice for the symbol update each week.

Yes, if you are workng primarily in WL 5, WL 5 is much faster in its ability to directly update its data set(s), and so no need to use WL 4.5 data set. But if you are going to be using both WL 4.5 and 5, you might as well create it in WL 4.5 and update 4.5, and just have WL 5 link to 4.5 data without having to update 5.

For WL 4.5's data set, it is well labeled "Modify" so click that and hit Control A to select all, hit Control V to paste that which you have already copied from elsewhere and new symbols will populate the data set.

For WL 5, Control C (copy) and Control V (paste) do not work. Of course you can use Control C to copy updated symbols from elsewhere, but then in the data set, after hiting Control A (select All), hit delete, then hit the right mouse button then choose paste from the pop up menu, and new symbol set is populated into data set; then hit Apply Symbol Changes.
profile picture

Eugene

#14
QUOTE:
Because of various aspects of WL5, I think "beginners and non programmers" probably should still stick to WL 4.5 for now for the heavy lifting.


Beginners will benefit from climbing on the WL5 bandwagon which will cut their learning curve halfway (just by eliminating the need to know two languages in the future -- Pascal and C#). Rule-based strategies in WL5 also cater for non-programmers.

QUOTE:
WL 5, at this moment, cannot convert and cannot run any of the many hundreds if not thousands of public Chartscripts created prior to WL 5.


While some see it as a problem, honestly, it's somewhat bloated. Several hundred chartscripts are just variations of one primary idea, another several hundred are either untradeable or questionable by their concept and so on. Truly robust scripts tend to be short and concise in contrast to some scripts so complicated that it makes them hard to grasp and explain in plain English.

What won't translate to WL5 is "trading the equity curve" (by design, unfortunately) and script-controlled position sizing (until it's implemented).

QUOTE:
For WL 5, Control C (copy) and Control V (paste) do not work.


This is (already) fixed in 5.1. Meantime, regular Windows shortcuts Ctrl-Insert and Shift-Insert will do the same job.
profile picture

swuzy

#15
Quite True That WL 5's Wizard Is Far More Capable Than WL 4.5's.

Especially important for nonprogrammers, it is quite true and most important that WL 5's wizard is much more flexible and powerful in ability to access and use all varieties of listed technical indicators. WL 4.5's wizard is comparatively very limited in flexibility and ability to use only a much smaller number of indicators. So if you have some knowledge of technical analysis, you will have much more satisfactory and much more powerful tools in WL 5 to create, and try out your own trading stategies.

Hopefully WL 5 will soon have access to public stategies.

Because I am not a programmer and have not studied the programming aspects of WL, I often like to rely on public chartscripts of WL 4 to study how some (relatively simple) things are done, and try to copy and paste segments into my combo scripts.

At this time, I do not see any repositary of public Strategies for WL 5 or module to access such public Strategies. May I assume this will be available in 5.1?

Ability to Sort and Rank in Alert Window and Performance Visualizers

At the moment, after running a WL 5 Strategy that uses the Signal Name column to rank and prioritize Alerts, in the Alerts window, the Signal Name column does not sort correctly, and I cannot Control C copy and paste to an Excel sheet to try to sort in correct manner.

For me, this very simple shortcoming of WL 5, an inability to visually prioritize symbol alerts for further review and trading, is a fatal flaw that prevents me from bothering to use it to generate trading alert signals.

So for people whose Strategies and Watchlist combination does not generate a lot of Alerts, WL 5's very superior Wizard is clearly the way to go.

New Sortable Performance Visualizers Coming in WL 5 Extensions

A fix is coming to allow for sortable Alerts. I am looking forward to it. I am not sure if WL 5.0 has an Extension Manager (I do not see it under Tools), or maybe it is actually a 5.1 item.



profile picture

Eugene

#16
QUOTE:
At this time, I do not see any repositary of public Strategies for WL 5 or module to access such public Strategies. May I assume this will be available in 5.1?


There is a "repository" in the making: Community.Strategies.

And the module is called the "Extension Manager". Working with the EM is going to be very much like Firefox handles its Extensions: everything is going to be easy, descriptive and hassle-free. You'll go to the new WL site to download new addons, and the EM will alert you on startup when an update is available for some Extension.

QUOTE:
and I cannot Control C copy and paste to an Excel sheet to try to sort in correct manner.


WL 5.1 already has copying/printing nicely implemented, all we have to do is wait for its release.

QUOTE:
At the moment, after running a WL 5 Strategy that uses the Signal Name column to rank and prioritize Alerts, in the Alerts window, the Signal Name column does not sort correctly,


I understand. Please revisit this page for more details as I've already coded a prototype of Performance Visualizer with sortable Priority column. Hence there will be no need to convert priority into string and sort it.

QUOTE:
A fix is coming to allow for sortable Alerts.


Sorry I can't recollect if this ever was a problem in 5.0 but having just checked in 5.1 to find the columns of the Alerts PV correctly sortable.

QUOTE:
I am not sure if WL 5.0 has an Extension Manager (I do not see it under Tools), or maybe it is actually a 5.1 item.


It's an inherent feature of Wealth-Lab Developer 5.1.
profile picture

fsimms

#17
Danny, I am starting to trade option credit spreads with WealthLab systems. I am new to this so I haven't a had a record to verify that it will work, but I do have a lot of confidence in the potential.

Don't look at how a system does vs buy and hold with option spreads. You just want a high win rate. The spread will work as your stoploss system. If you win 75% of the spreads then the 25% that you lose will not hurt you if you handle the spread properly and have a good risk/reward on the spread.

I am looking at divergence systems to give me good signals to use on spreads. One example is below.

http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=4305

Bob