Builder continues to blow my mind

[David Staheli]

Besides being an outstanding strategy builder, Adaptrade Builder is an absolute masterpiece in software development. What an incredible accomplishment!

[MikeBryant]

Thank you. ...I swear, I didn't pay him to say that.

[Mohit Gupta]

Can you please share any basic strategies and do these work in live trading?

On Wednesday, July 3, 2024, David Staheli <da...@staheli.org> wrote:

Besides being an outstanding strategy builder, Adaptrade Builder is an absolute masterpiece in software development. What an incredible accomplishment!

--
You received this message because you are subscribed to the Google Groups "Adaptrade Builder" group.
To unsubscribe from this group and stop receiving emails from it, send an email to adaptrade-builder+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/26d06ee2-e074-433e-bf59-a12505f85845n%40googlegroups.com.

[C G]

The software opereration and interface are excellent. 

But,  one issue with auto builders like Adaptrade is curve fitting. 

All of the strategies I've developed using the software fail when exposed to blind data... or data not in the software.

I'd like to know if anyone has found truly amazing strategies that have been profitable for 3 months or more... with real money. 

Thanks. 

On Sat, Jul 6, 2024, 2:19 AM Mohit Gupta <mohit...@gmail.com> wrote:

Can you please share any basic strategies and do these work in live trading?

On Wednesday, July 3, 2024, David Staheli <da...@staheli.org> wrote:

Besides being an outstanding strategy builder, Adaptrade Builder is an absolute masterpiece in software development. What an incredible accomplishment!

--
You received this message because you are subscribed to the Google Groups "Adaptrade Builder" group.

To unsubscribe from this group and stop receiving emails from it, send an email to adaptrade-buil...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/26d06ee2-e074-433e-bf59-a12505f85845n%40googlegroups.com.

--

You received this message because you are subscribed to the Google Groups "Adaptrade Builder" group.

To unsubscribe from this group and stop receiving emails from it, send an email to adaptrade-buil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/CAMeecZiUG0LmvQpRhF%2B67jasK34cntgdP%3DGFWYT45Wz7ZejKYA%40mail.gmail.com.

[rahanq]

Hi guys

Hope you are well

Same here, lovely piece of software but I just cannot build a system that works out of sample 

Maybe it’s me and I am making mistakes, if anyone has built some winning systems, and can show me how, I’d love to learn 

Best wishes

Sent from my iPhone

On 7 Jul 2024, at 13:49, C G <cgano...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/CAC_WV9jqN-8mS%2B0s9XunpbC9ypK5R6yCHpCwgb2wariZdLvkFg%40mail.gmail.com.

[Hugh H]

You should always test your strategies on unseen out-of-sample data.  All strategy builder products are built on in-sample data. 

Hugh

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/CAC_WV9jqN-8mS%2B0s9XunpbC9ypK5R6yCHpCwgb2wariZdLvkFg%40mail.gmail.com.

[matt n]

“Same here, lovely piece of software but I just cannot build a system that works out of sample”

Common issue with trading strategies. Even the big prop funds deal with this at times. 

Matt

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/9F033D3C-FBAF-4990-A097-D8FD83910E43%40googlemail.com.

[Jeff Kirkson]

Hi everybody,

 

With Builder I can build hundreds and hundreds of strategies.

If I test them 3 months later, 60-70% of these strategies have lost money but there are a good 30-40% of those that have earned it.

The biggest difficulty I have is selecting the right strategies, those strategies that earn real money in the real market.

These strategies exist among those built by the builder, maybe they don't last a long time, but they exist.

I have tried many selection methods (Montecarlo analysis, walk forward, stress test, double OOS period, shape of the equity curve,

selections based on metrics, reverse engineering methods etc.) but I still haven't found a method that can give me a sufficient guarantee

in terms of the probability of choosing profitable strategies and avoiding those losers.

Evidently there is no such clear parameter that guarantees the strategies persistence or the positive correlation between the past and the future.

So today I have to say that the builder is an exceptional software for creating (even) winning strategies, but choosing profitable ones is a completely different matter.

Have any of you overcome this problem?

Thank you.

 

Jeff

 

Da: 'rahanq' via Adaptrade Builder <adaptrad...@googlegroups.com>
Inviato: domenica 7 luglio 2024 17:10
A: adaptrad...@googlegroups.com
Oggetto: Re: Builder continues to blow my mind

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/9F033D3C-FBAF-4990-A097-D8FD83910E43%40googlemail.com.

[C G]

Thanks Jeff. 

How many strategies are you trading?

Are you sym trading or using real money?

The challenge is that if you develop 10 strategies... 3 out of 10 are profitable...7 are not. So you probably lost money...assuming you traded with the same exposure... say 1 contract and the same vehicle... let's say ES.

Then you turn off the 7 that didn't work... and left the 3 on that did. 

But you're still rolling dice as the 3 you left on could very likely turn south and lose money, too. 

So... to me... this is not safe for your money... or mine. 😆

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/0c8b01dad08f%24649a3760%242dcea620%24%40gmail.com.

[MikeBryant]

To those who've read my newsletter articles, I probably sound like a broken record, but it's important to understand the difference between curve-fitting and over-fitting. In fact, I would say that's the most important factor in developing trading strategies, regardless of the method or tool you use to do it. Curve-fitting is just another name for optimization. You have to choose values for strategy parameters somehow, and if you're not optimizing, you're just throwing darts at the board. So optimization or curve-fitting is just a systematic way to choose parameter values. There's nothing wrong with that. The problem is over-fitting, which means choosing values for your parameters by fitting the strategy to the noise rather than the signal. If you fit the noise, the strategy won't generalize, which means it won't hold up out-of-sample.

I won't go through all the methods in Builder that are designed to help you avoid fitting the noise, but I will comment on one mistake I often see: traders who have unrealistic expectations and try to force more from the market than the market is capable of giving. In part, I blame this tendency on unscrupulous strategy vendors who knowingly hawk over-fit strategies or who had one successful trading period with real money and imply the method they successfully used for that relatively short period will hold up indefinitely. The constant barrage of promotions along those lines can create unrealistic expectations.

If you're looking for a better overall approach for developing and implementing systematic trading approaches, here it is: trade a diversified portfolio of trading strategies, continuously develop new strategies, and regularly replace under-performing strategies with new ones. As long as you keep your expectations in check and use the features in Builder to minimize over-fitting, that's probably the best overall approach I can suggest.

[Tony Holland]

In my opinion all strategies fail over time but adapt lets your build new ones all the time to fit the environment.  I have built one on gold that has performed for years and using longer time frames yields better longevity.

Mike has done a study on building one every day that is an interesting concept.

I tried it for a few months but did not keep up with it.

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/CAC_WV9hd0eK%3D1E%2BQWjv4krW2t7RNs%2BcRQbx%2BRffNNHRSADM9Vw%40mail.gmail.com.

--

Tony Holland

502.664.2820

[MikeBryant]

It's not in-sample if it's not used in the building process.

[Hugh H]

A friend of mine posted this on Friday.

The thing is, you can never rely on a performance report or backtest. Backtests are not reality. :) They are estimates of the past. They provide some information, but the only reliable information is a live-money trading record.

Here is the level of certainty when it comes to trading reports. Number 1 is the best (It's not a backtest!), and number 4 is the worst.
 

1. Real money living trading record
2. Unseen walk-forward testing
3. Out-of-sample backtest
4. In-sample backtest
5. 
   

Hugh

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/ea071693-6be7-4f45-a36f-09bfef854d81n%40googlegroups.com.

[MikeBryant]

Absolutely. That's why almost all my newsletter articles -- at least the ones that include strategy results -- end with a reminder to test forward in real time.

[C G]

Yes... this is basic strategy development.

It's important for users to understand the time and energy involved with using Builder. I have spent hundreds of hours testing strategies with Builder... and I don't have any that work. Every one I check with blind data... failed. I'm out of time. 

One can spend thousands of hours and build thousands of strategies with Builder... and not know if anything works.

Then you have to manually check the thousands of strategies to see if any of them work with blind out of sample data. This will take thousands of hours. 

The point of this follow up to the thread "Builder continues to blow my mind"...is that I immediately thought this user may not fully understand that the strategies that are being created... fail with real money. So... this is just a warning that... strategy development apps can blow your mind... and also all your money. 🙂

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/1941009b-56ac-4b7b-83ec-22b54adee0dcn%40googlegroups.com.

[Hugh H]

For me 95 out of 100 strategies  fail ....if it was easy money everyone would be doing it!

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/CAC_WV9h5gzybVwj6CHb3iRiZA4Ss-ga_YsDxc2szs6dKymhyBQ%40mail.gmail.com.

[MikeBryant]

When you decide to trade the markets, you're basically deciding to be a hedge fund trader or CTA with your own money. It's not for everyone.

I would also add that if you feel you need to check thousands of strategies after running Builder, you're not using the tool correctly.

[C G]

I just wanted to be transparent that I found no strategies that work with real money. 

I read your tutorials and even took training from one of your vendors. 

I have heard the same same sentiment from other people. 

To view this discussion on the web visit https://groups.google.com/d/msgid/adaptrade-builder/31f7d32d-97c5-43fd-a144-5b207f5c4789n%40googlegroups.com.

[MikeBryant]

I'm sorry to hear that. Ultimately, Builder is a tool to automate and streamline the strategy building process. How the tool is used and what you get from it is up to the user.