The Holy Grail Trading System: Backtest (editied to remove sarcasm)

This is a revised edition of this post.

The previous post text was meant to be sarcastic and not intended to be an attack or flame war.

It took a lot of work and time to develop this system. I hope it was not a waste of time.

Images reprinted with permission, Futures Truth Magazine

A red flag for me on a quick glance is 635 trades over 18 years with ZERO LOSSES! It would be good to see the rest of the TS Performance Report (drawdown info, etc.) as well as the starting capital. But 635 consecutive trades with zero losses invokes a “too good to be true” first impression. Is this system live and trading on C2?

Indeed, a backtest with Trade Number on the X-axis can hide drawdowns completely. There could well be horrific intratrade drawdowns that were simply held for weeks or months until they became winners, and these drawdowns are not visible unless the equity curve is shown on a Timeline. It’s not too difficult to construct an inverse volatility strategy with such an equity curve… when viewed in Trade Number space instead of on a timeline.

1 Like

Why drawdowns info is hidden in the report?

Yes … this seems to be the MO of many high win% systems. Also, the starting capital amount is not shown and obviously turning $1M into $1.4M over 18 years isn’t impressive (1.9% CAGR), turning $100K into $1.4M is very good (15.8% CAGR), and turning $10K into $1.4M is fantastic (31.6% CAGR). But zero losses in 635 trades certainly warrants more explanation. As the old saying goes … extraordinary claims require extraordinary evidence (especially if soliciting money for investment in some way).

Relatedly, here’s a blog post from jonathan kinlay about such magnificent equity curve strategies and what they can hide.

There is also the ability to unwittingly create beautiful equity curves in Tradestation’s Easylanguage due to quirks in their trade engine and rules if you are not familiar with them. When I was new to Easylanguage and trying to build my own strategies, I landed on what appeared to be a can’t lose system on daily bars, but it also looked too good to be true. Upon closer inspection I noticed that every day a trailing stop would close out the trade, but the stops were always relative to the absolute high of the day and the smaller the trailing stop, the better the profits.Win % was only about 45% so I thought it might be legit, but doing “inside the bar backtesting” crashed it completely. So it wasn’t real, although there were no errors in the Easylanguage code or basic idea … just a quirk of how the TS trade engine worked and my ignorance at the time of the need to do inside the bar back testing when using daily bars.

I’m not saying anything like that is happening here, but I’ve never seen a valid system that could make 635 consecutive trades with zero losses, and it is legitimate to be skeptical of that kind of record. Either it is a “hang on forever until the trade turns around” system, or indeed a holy grail. I’d like to see more details.

1 Like

Yup, I’ve been there, although not in TS personally. Even when a developer is being careful not to be fooled mechanically (ie, peeking into the near-term future such as you describe), it’s all too easy to find a strategy too closely optimized on the ever-present noise in market data and believe it’s a discovery of a tradable edge. Statistics alone cannot fully prove a strategy’s future value.

Been there as well. Have been developing strategies in TS for over 25 years (since the beginning). Have found several “Holy Grail” systems over the years with even better back test results than the one shown here by using quirky, missunderstood tools such as Renko bars, Kagi bars, combined with the ever so treacherous Percent trailing profit stops and much more. These things can’t be evaluated intrabar and so you never know how the actual prices moved, you only see the high and low of each bar. Many, many pitfalls in EasyLanguage strategy developement and I have yet to find a “Holy Grail” system that actually works well in reality.

1 Like

No such thing as a “Holy Grail” system. Every system has a weakness.

1 Like

As the disclaimer at the bottom of the equity curve states, this is a curve fitting strategy that benefits from hindsight. The curve will start to go down as soon as live trading begins, or at best will be random.

Are you referring this backtest with Lighthouse model or a different model? The most important is out of sample from backtest. I have seen many great backtest but when apply to real account, the return and DD are huge different from backtest performance. In addition, some strategies in the real account can perform well in one direction only, but when the market is changing, the strategy doesn’t work very well. I personally don’t relly only backtest, the model must be test in the real trading and in longer time periods. You should include what is DD in your backtest.

1 Like

Please realize, everyone, that out-of-sample testing essentially becomes the equivalent of in-sample testing…if it’s done repeatedly to optimize an algorithm.

Lighthouse is the antithesis of the “Holy Grail Trading System” because it was very carefully designed with many years of experience. In developing Lighthouse, I have applied a lot of hard lessons learned such as the points made in the replies above about over-fitting the system to the data and hiding undesirable and risky characteristics in an incomplete performance report of summary statistics.

Lighthouse is running live on C2 to demonstrate that it is properly developed, robust and not over-fit to the data. I am generally against publishing performance reports because they are so often wrong, bad or just plainly manipulations.

The “Holy Grail Trading System” is indeed too good to be true. The disclaimer tells the whole story. The name is also a dead giveaway. The disclaimer at the bottom of the images also clearly states that this is not to be trusted or believed as a trading system that can or will work. It is bunk and junk except…

…who bothers to read the disclaimers anymore?

The reason why I posted “Holy Grail Trading System” performance reports here is because people keep asking me for to provide performance reports for my systems. Having third-party audited or independently tracked systems that run live in real-time is the most honest, trustworthy and transparent way to display trading system performance.

Unfortunately, there is no way to be sure that a good, great, or even an excellent trading system performance report will represent future performance because even the best trading systems can get out of sync with market conditions. Some systems even burn out while others fade away.

The naive and the untrustworthy can always hide behind the boilerplate disclaimers.

The “Holy Grail Trading System” was not meant to deceive or mislead. It is an educational tool to help us learn to be weary of unqualified performance reports. Unqualified, in this context, means historical and hypothetical performance reports that do not completely detail the development, testing and validation methods in a totally transparent manner.

Who is going to give out that kind of information? Have you ever seen that level of transparency in any trading system performance report? Would you trust such a disclosure of that type of detailed methodology? How many traders would even know how to analyze such information?

I really do hope that we can all learn form this exercise. A lot of time and effort went into picking all of the tops and bottoms on the price chart and coding the strategy to stop and reverse at every pivot point even near weekends and holidays.

Thanks to Futures Truth Magazine for allowing me to reprint material from the “False Profits: Exposing Bias with Independent Performance Audits” article series. The series closing article displays these graphics:



So in other words, it was worded that way for click baiting?

That is exactly what it is!:money_mouth_face:

My motivation is to expose how easily it is to fall into the emotional traps and succumb to manipulation when we are looking for trading systems and profits.

I made all of the worst mistakes as part of my trading education (paying the dues, as they say).

It is generally accepted that about 90% of all traders fail and quit trading within five years of starting.

I would not have learned very much if I would have quit in 2002. I am still learning every day and hoping to keep it that way.

Smart people learn from their own mistakes. Wise people learn from the mistakes of others.

Maybe it was a mistake to post this “click bait.”

If it helps others learn to slow down and think before “throwing good money at bad” then it will be worth whatever criticism comes as a consequence.

I hope to learn something from it. Thank you for teaching me a lesson today.

“Live long and prosper” _\V/


Thanks for the link!

That is a good article.