My 2 Cents

[AdaptiveTrader]

1. TSL doesn't work. Tried it personally. Was not able to develop a single working strategy.

2. Price Action Lab - Are you kidding me for $8000 lifetime or $3000/year.

3. Can we please focus on what would really improve Builder (e.g. portfolio support, more adaptive indicators, etc.)

I personally think it is highly disrespectful to Mike and wasteful of this forum's time to spend this inordinate amount of time on other platforms. I want Builder to improve as well but not by making into a clone of another product or shoehorn features that don't fit. I am currently running severing EOD and 60 minute strategies developed with Builder. My sense is the short time frames are much harder as there is so much noise

[enz]

AB should be able to perform WFA as well as testing and eventually modifying already existing trading system formulas, not only "inventing" new ones. The best start is always an existing trading system which have worked fine in the last years and may need to be improved/walk forwarded over time.

[AdaptiveTrader]

WFA sounds good in principle but I think is at odds with Builder's philosophy. WFA simple selects new parameters after the WF period has ended. Why not simply reoptimize/rebuild your strategy after a period of time (such as the OOS period)? Every feature you add makes Builder that much more complicated and difficult to maintain.

I disagree with having Builder start with an existing system Why would you want to bias Builder with ideas about the market? Why not let Builder discover them?

[Rick]

On Mar 19, 4:12 am, AdaptiveTrader <collison...@gmail.com> wrote:
> 1. TSL doesn't work. Tried it personally. Was not able to develop a single
> working strategy.

Are you sure you used it properly? Did you try it at work or purchased
your own copy?

> 2. Price Action Lab - Are you kidding me for $8000 lifetime or $3000/year.

It is $3000 for the first year and then you pay the difference if you
want to get a lifetime license with no restrictions. You can run 32
copies on 8 quad-core machines with no restrictions. I'm setting up
something along these lines currently. I will have this program do
data mining using 30, 60, 4H and EOD data of most liquid securities
US, UK, Aussie and some Asian markets. I am also looking for ways to
use the Price Action Lab output with AB.

> 3. Can we please focus on what would really improve Builder (e.g. portfolio
> support, more adaptive indicators, etc.)
>
> I personally think it is highly disrespectful to Mike and wasteful of this
> forum's time to spend this inordinate amount of time on other platforms. I
> want Builder to improve as well but not by making into a clone of another
> product or shoehorn features that don't fit. I am currently running
> severing EOD and 60 minute strategies developed with Builder. My sense is
> the short time frames are much harder as there is so much noise

I think it is always good to know and discuss what others do. It is
not disrespectful IMO.

[AdaptiveTrader]

Yes I used it correctly. Tried for 8 months even worked with developer.

Can you explain getting the output of PAL into Builder?

[Rick]

On Mar 19, 6:57 pm, AdaptiveTrader <collison...@gmail.com> wrote:
> Yes I used it correctly. Tried for 8 months even worked with developer.
>
> Can you explain getting the output of PAL into Builder?

PAL can output code in a .txt file. See their version release notes.
Basically, just the conditions and parameters of patterns are listed
in the file. The idea is to use this file as input to a GP engine to
develop strategies. You can add a few indicators to act as filters,
like a fast and a slow MA, MACD, RSI, etc.. I see this as a very
promising approach. You start with something basic that works and it
is improved at a higher level. You don't start with random choices and
then optimize them like TSL and AB do. I will have to first use PAL to
generate as many files as possible. Then I will use AB or, if I can't,
another GP engine to work with the files. Finding ways to keep
busy...:)

> > not disrespectful IMO.- Hide quoted text -
>
> - Show quoted text -

[mandelmus]

1a.  With all those "super duper extra special features", you couldn't get a single working strategy?  No lo creo!  Why do you think TSL did not work?  But, I'm not entirely surprised, anyone who claims their software is "the best in the world and no one else comes close" may be slightly delusional.

 

1b. How long ago did you try Trading System Lab (TSL)?

 

1c. My understanding is that he doesn't offer a trial; how much did he charge you to try TSL?  Or, did you work out some other kind of arrangement?

 

2. Ya, Price Action Lab (PAL) does seem a bit expensive for software that simply looks for specific price patterns ... kind of like a one-trick pony.  But, it seems some people are finding success with it.  If it produces any edge-tested non-correlated strategies, I'd consider applying them in some type of acrary-style (it's an interesting approach and if you have the time to read through it all) and/or algodude-style strategy.

3a. I think "portfolio support" would be a really great feature, doesn't TSL already have that? I like some of the portfolio optimization features used in Optimal Trader

 

3b. What does "more adaptive indicators" mean?

 

You should not interpret our discussions and suggestions as disrespectful.  If Mike doesn't like our suggestions, he'll simply ignore them -- nothing really lost in posting the suggestions.  But, I feel it's better for us to post our suggestions so Mike can gauge which issues are important to us (refer to this article on why feature requests get ignored and/or take so long to implement).

 

As I mentioned in a previous thread, if you don't like my posts -- ignore them!  Otherwise, I believe there is tremendous benefit in discussing the advantages and disadvantages of alternative software.  Discussions about alternative software on this forum should not be interpreted as an attack on Builder.   I consider our group to be highly intelligent traders who understand the benefits (or, potential benefits) of using artificial intelligence type software.  I'm sure most of us continually look to improve our arsenal of trading tools.  I'd rather discuss artificial intelligence type software on this forum, with other like-minded traders like yourself, than on some random thread on EliteTrader, TradeStation forums, Trade2Win, etc.

 

I don't see why Builder shouldn't grow to its fullest potential.  If that means looking like a "clone" of some other software, so be it.  I've also developed functional strategies in Builder, but that doesn't mean I should stop asking for more features.  Just recently, I found a way to get many strategies that perform well down to the tick level.  It took me several months of reading and experimenting with different inputs to figure out how to get the results I wanted out of Builder.  I'm eager to share some of my settings, parameters and results with y'all soon.

[mandelmus]

AdaptiveTrader makes valid points.  WF optimization may not be beneficial for GP strategy development.  But, if you have a suggestion for how GP+WF optimization might work, please share it.

[mandelmus]

Have you already started data mining with PAL?

 

Uumm, how much are all those computers costing you?  Can you spare a few quad-core machines for your trading brothers? :D

[AdaptiveTrader]

Tried TSL on and off for about a year.

No trial versions.

Portfolio support yes but only small portfolios.

I don't have a problem comparing Builder to other products but trying to make Builder INTO one of these other products is a mistake. I know it's not yours or anyone else's intention but it feels that way. Neural nets, patterns from PAL, etc. Before you know it Builder will have hundreds of features that few people use. I'll say it again I think the focus should be on improving the core genetic engine, adaptive indicators, portfolio support, and better ease of use. I think some training in some form would also be valuable.

Most indicators have fixed lookback periods and or levels that indicate overbought/oversold conditions. Why choose a 14 period RSI for example and what happens when the market cycle is no longer 14 periods. What happens when the market hits a RSI level of 80 and stays there but your strategy says the market is oversold? The indicators provided with most technical analysis products were invented 25-30 years ago before computers were widely available. They are now in such wide spread use as to make them almost worthless. The market has become too efficient with respect to these indicators.

[Rick]

The hardware costs about 10K. The software costs 5K (PAL + misc.). I
had the idea and I approached a hedge fund. They have tried everything
in the past and concluded system trading/indicators/TA, etc. doesn't
work. I explained to them that PAL looks at price action and micro-
patterns. One of the partners liked the idea of using this program
with data files of thousands of securities with the possibility of
identifying high probability setups (they will do the stat tests). We
figured out a way of auto-creating PAL search workspaces by modifying
an existing one. This saves a lot of time. PAL search can do 100 files
with 100 different target and stops during a single search session (we
will use only 5 combinations of targets and stops). This comes to 3200
files per session (32 PCs X 100). Divide that by 5 (30, 60, 1H, 4H,
daily - data files) and that yields 640 different securities per
session. The full search takes about 4 hours per file but it varies.
Total estimated run time is 17 days (100 X 4/24) for doing 640
securities with 32 quad-core machines. This is only an estimate. We
will start next week after running a few tests. Be sure they will
recover the cost and make a profit too. They just sent a proposal to
hedge funds in Italy and Australia for repeating this for their local
markets for a hefty fee (pending approval from PAL owner because this
is against their license terms). I will let you know in about a month
of the first results.

> > use the Price Action Lab output with AB.- Hide quoted text -

[mandelmus]

so, you paid for TSL?  or, did he let you try it free?

 

don't fret, there's no way mike will add anywhere near all those features to Builder

 

one of TSL's demo videos shows his use of a set of indicators that calculate optimal cycle lengths to use as inputs to the regular indicators ... is that what you're talking about in regard to adaptive indicators?

[mandelmus]

so, what did they conclude worked?

[mandelmus]

Did you also spend $60,000 to try out TSL or did you work something else out?

 

Barna's demo videos have given me some great ideas and the features he says he is adding sound good.  However, based on reviews I've read so far, it's not worth $60,000.

On Tuesday, March 27, 2012 2:49:26 PM UTC-5, braincell wrote:

AdaptiveTrader wrote:
* 1. TSL doesn't work. Tried it personally. Was not able to develop a
single
 working strategy.

It might not work as advertised, but it works sometimes. It takes a
bit of effort to get to use it properly, but the performances also
aren't as great as his flash demos make them out to be. You can't say
it doesn't work though. Take note of the systems TSL has posted on
Futures Truth. I really doubt they would be posting false information,
even if those systems are just slightly modified versions of the Big
Blue 2 system which Mike Barna developed some time ago.

[Rick]

On Mar 27, 4:12 pm, mandelmus <gmb1...@gmail.com> wrote:
> Did you also spend $60,000 to try out TSL or did you work something else
> out?

Of course he hasn't tried it. He hasn't even tried a demo because it's
not available. So I don't know how he reached his conclusions. One
thing I just found out is that this guy is developing his own data
mining software:

http://www.elitetrader.com/vb/showthread.php?s=&postid=3460654#post3460654

"Offline data mining is one of the main reasons i became good at
assembler and i can tell you my applications are beating those used by
some hedge funds at least 2x in terms of speed - meaning you need 2x
as little hardware for the same results, and this is a time and cost
saving. I can probably improve that even further, but never mind
that."

His applications are beating those used by some hedge funds.:)

>
> Barna's demo videos have given me some great ideas and the features he says
> he is adding sound good.  However, based on reviews I've read so far, it's
> not worth $60,000.

Too much hype and no substance IMO. AB does the same work at a
fraction of the price.

>
>
>
> On Tuesday, March 27, 2012 2:49:26 PM UTC-5, braincell wrote:
> > AdaptiveTrader wrote:
> > * 1. TSL doesn't work. Tried it personally. Was not able to develop a
> > single
> >  working strategy.
>
> > It might not work as advertised, but it works sometimes. It takes a
> > bit of effort to get to use it properly, but the performances also
> > aren't as great as his flash demos make them out to be. You can't say
> > it doesn't work though. Take note of the systems TSL has posted on
> > Futures Truth. I really doubt they would be posting false information,
> > even if those systems are just slightly modified versions of the Big

> > Blue 2 system which Mike Barna developed some time ago.- Hide quoted text -