ETMA_The Evolution

Interesting.

This one goes for u C.

Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance
David H. Bailey, Jonathan M. Borwein, Marcos López de Prado, and Qiji Jim Zhu
http://www.ams.org/notices/201405/rn...=244+272699400

Thanks for this Bass. I hope all is well mate.

Laurent Bernut, ex-Fidelity short seller, ex-Hedge Funds analyst, ex-CPA, algo trader

Now, let’s shake some cerebral tofu here. Regardless of the asset class, there are only two strategies: mean reversion and trend following.
mean reversion:
Pros:

1. high win rate
2. Compounding: small profits compound quick and increase the Capital balance
3. attractive to investors: who has not been conned into Short OTM short options, fundamentals pairs trading?

Cons:

1. left tail: the captain of the Titanic had a 99% win rate until…
2. Ugly tail ratio: right tail is short, left tail is deep, good recipe for bad news
3. Long period of recovery: picking pennies in front of a steam roller

Trend Following
Pros:

1. Robustness: trend following endures. Despite all the news on trend following being dead, trends are and as long as markets go somewhere, trend followers will be the humble valets
2. Simplicity: complexity is a form of laziness. Academics, double shot, soy latte, go fetch
3. capital appreciation: we all are trend followers, earnings momentum, newsflow, fundamentals rate of change etc. This is not compounding. Buy & Hold is trend following
4. Solid risk management: trend followers are portable risk managers

Cons:

1. Lumpy P&L: now that is a real kill joy. Trends are like princess charming. You have to kiss a lot of frogs and your P&L will erode until then
2. Correlation: trend following 2.0 advocates for benchmarking, 1.0 is absolute which works with diversified asset classes such as futures

Hybrid
Trend following with mean reversion. The idea is to buy on weakness along the dominant trend and scale out, mean reversion. This is what we have been doing. It works
Pros:

1. high win rate: that comes with the quick compounding of mean reversion
2. capital appreciation: this comes from trend following

Cons:

1. Counter-trends: in counter-trends, trend followers get either shaken out, or stop trading, hence the drawdown
2. FIFO: this is an idiotic administrative rule, whereby scaling out is depleted from original trade. You need LIFO, last trade scaled out independently of pre-existing positions. That is a big one, not good

Tofu Shake
Now that i have your attention, how about this: suppose you can manage all positions independently, how about:

1. trend following with mean reversion AND
2. Mean reversion with trend following, contained in the same strategy

Let me explain. We have the hybrid under control: trend following with a bit of mean reversion. we have been in real money production and it produces. We enter Long on weakness and scale out when trend weakens. we have however been puzzled by the money left on the table. There is room for mean reversion within that. So, we came up with a robustly simple algo that exploits mean reversion (with a trend following component) that does not conflict with the overall trend following. Here is the concept
Mean reversion is essentially is a time frame issue. On longer and shorter time frame, markets trend. Yet at our trading periodicity, it sometimes mean reverts.
So, we have used this mean reversion to enter and exit within the mean reversion horizon. The catches are:

1. Partial exit on mean reversion: the idea is not to close 100% of the trade, but a portion of it. The remainder is closed upon a swing.
2. Trade the dominant: one of the tenets of mean reversion is de-trending. Lazy. Trade with dominant trend as it moves up or down in one direction. Then, when counter-trend hits, trade the counter-trend side only

The objective is to cover risk through mean reversion and let the remainder generate the profit through trend following.
Rules are simple, coding a bit hard as it involves multiple positions on both sides.
Pros:

1. Accelerated compounding
2. High win rate: trend is your friend
3. Hedging on counter trends: you will swing trade counter-trends, generate profit and hedge

Cons:

1. Good luck programming this. So far we are launching the mean reversion with trend following next week. It fits nicely into our trend following with mean reversion. Mean reversion will go into real money production in 2–3 weeks. i am ordering a bigger wine fridge. Counter trend mean reversion will have to wait few more week, maybe months. Practical problems abound. We have narrowed it down to a tit-for-tat money management. that is the most robust under the circumstances and according to AI game theory, it seems to work
2. Good luck finding a jurisdiction and a broker that will allow independent order management
3. Indicators don’t work well. This is pure logic

This approach solves a lot of problems inherent with either strategy. It however pre-supposes symmetry, unambiguous rules. This exercise forced us to simplify our rules even further. Our goal conditions our architecture. Our goal is synchronicity: same strategy across multiple time frames. Now imagine this with trend following + mean reversion && mean reversion + trend following .

Happy to help and collaborate with those who are interested. Apart from synchronicity, i can’t think of any more interesting project in the quants space. It flips everything You have thought you knew upside down, fascinating .

Hi Bass,

the tofu is an intersting approach and close to what I am looking for.....but far from automated and just demo
Price is the only indicator needed, the rest is distraction
A broker in Switzerland should be able to offer you independent order management
However, at the end of each day it makes sense to merge long / short positions to minimize rollover costs
The only way to limit tail risk in my opinion is a significant capital base combined with very little leverage
I am pretty sure fix parameters won't do the job in the long run.... you have to be highly adaptive in regards to time frame, vola, market conditions, pattern recognition, price interpretation... etc.

How close is such a model to what a market maker and liquidity provider is doing?

Would be cool to hear about your progress.

Good luck, Hudi

Hi Bass,

I disagree with all the trending following with a mean reversion; or toufu; whatever it is called.

at the end of the day; you are trading 2 strategies; trend following; and mean reversion; and the outcome will be the sum of both PNLs.

therefore; you are better off keeping the two separate; and trade them independently of each other. It saves you trouble in coding and managing it over the long run.

It is always nice to have a portfolio of various strategies that are not correlated. You will need to allocate capital accordingly.

am talking from my own experience.

Its Being 2 years that i didn't post in my thread because i haven't taking a trade since January 2018, people still subscribing to me for the high quality researching that i did and still doing in my spare time which is increasing these days.

Several subscribers are still asking me questions in Private, this is the latest :

Hi Boss
Can you teach me how to position my stop loss in scalping forex?
(Obviously a newbie trying to conquer the universe with a scalping strategy, welcome to the grind ...)

Buddy, I've got one answer for you : Test until you die OR until you find a good back and forward tested level to place your stop loss.

Anyone that tell you otherwise will be bull*****ing you .

You are welcome in advance because i know this tip will save you a lot and lot of money, trust me, 'being there - done that' and still doing it .

Almost anybody can make up a list of rules that are 80% as good as what we taught our people, “I always say that you could publish my trading rules in the newspaper and no one would follow them. What they couldn’t do is give them the confidence to stick to those rules even when things are going bad.”
-The Legendary Richard Dennis-

The funny thing is he is not even subscribed to me to get an answer to his question, hehe

Place it here for example ( Every strategy got to be tested extensively before using it live ) .

This is what trading really is, keep coming back to his wisdom again and again and again, timeless, its a shame that he is retiring, i wish i can copy all his trading knowledge brain on a USB stick and pasted on mine, always great to read, hear and see him teach or trade .

Best of luck on your retirement.

Doom and Gloom

When we are lucky and the Markets trends :

And when we are not and the Markets Bend :

Systematic Investor Series ft Richard Brennan – July 25th, 2021
Aka Copernicus