Momersion Indicator by JayJayD

JayJayD INVESTOR

| Reward Discussion

Hi gang,

Here's a new indicator, the Momersion, I found it in Michael Harris Price Action Lab blog.

In Harris words:

The day momentum died, mean-reversion was born. Now they have both merged into what I call momersion.

And respect to the Indicator itself:

The Momersion indicator measures momentum and mean-reversion over a specified period

In the main algorithm, I coded the two rules described in the Momersion post.

My idea is to use this Momersion as an auxiliary indicator, maybe to turn on/off some strategies.

The blog is really interesting, a friend told me once that he's following the series like a housewife watching soaps :D ; and I do it too!

Cheers, JJ

Update Backtest

JayJayD

Accepted Answer

Update Backtest

Notebook

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.

Christoffer Sí¸rensen

1.1k Pro ,

Thanks, Juan.

This indeed looks interesting, although I cannot see your code on mobile.

I can use this myself in my systems.

I also got more profits from mean reversion than momentum in 2014 and 2015, so I guess my hunch was correct.

C

Jared Broad

Pro ,

You're famous! Called out by the author! :)

Credit for the implementation goes to Juan José D'Ambrosio, an economist and quant trader. #momentum #trading #trade https://t.co/GXjKd2g9zo
— Michael Harris (@mikeharrisNY) October 1, 2015

JayJayD

Pro ,

@Christoffer, have you seen the Jim Simmons Ted interview? There, Simmons said that trend-following isn't working any more.

@Jared, I introduced a kind of early readiness in the code. The indicator have two periods, the full period and a minimum period, once reached the min period, the indicator is ready and returns a estimated value; and in the meantime continues growing until it reaches the full period. I don't really know if is a useful idea, but seems to fit in this particular indicator if you want to use it for intraday frames. What do you think?

BTW, the indicator is tested vs Excel and Python implementation... maybe you want to include in LEAN?

Cheers, JJ

P.S.: I need some vacation, I'm really tired of the hordes asking for autographs :o)

Christoffer Sí¸rensen

1.1k Pro ,

@Juan

I just watched the interview, it was good.

Regarding trend following: there was just a trend in China 2014-2015, so I guess not all trend following has disappeared.

I coded the momersion indicator for my python code at home where I calculate values for various ETFs and indices. It seems the NASDAQ index is in momentum phase now.

One idea one could backtest using the momersion indicator is to dynamically add and remove ETFs in the lists that ETF rotation strategies use.

JayJayD

Pro ,

@Christoffer, your last comment is great!

There was a some discussion last week about if the Momentum strategies are death or not.

Quoting again a post from Price Action Lab:

The argument was not whether someone today can or cannot find a trend-following system that will work. This is possible and such systems may exist. The argument was about changing market conditions that rendered short-term trend following unprofitable. Jim Simons thinks that this type of short-term trend-following is suitable for funds. I also agree because higher lags introduce higher risks.

Maybe that working trend-system more easily find in others markets, like China.

Thanks for sharing your thoughts!

JJ

Christoffer Sí¸rensen

1.1k Pro ,

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1088861

"Six Decades of Significant Autocorrelation in the U.S. Stock Market"

interesting paper.

Also shows the trendfollowing aspect has disappeared in recent times.

JayJayD

Pro ,

A picture Is worth a thousand words; a plot in our case :)

The first plot of the figure 3 (page 9) speaks for itself.

Thanks @Christoffer!

Jared Broad

Pro ,

@Juan - absolutely please submit to LEAN!

I like the slow-fast idea, but it might work better with orthogonal additional variables (non-price). Say volume, overall market trend (of a basket/S&P500), or sentiment (VIX stability). Its easier to solve n-dimensional problems with more variables.

IMHO basically all AI strategies are an automatic way of subbing in variables as they become relevant.

JayJayD

Pro ,

@Jared, I made the pull request with the indicator and the test.

Here is a notebook with the Python implementation used to test the results. @Christoffer, maybe you can check it, is a vectorized estimation.

Cheers!

Christoffer Sí¸rensen

1.1k Pro ,

I tested this with my mean reversion algorithm.

When I filter out trades with momersion >= 50, then I get improved results for symbols that can get into momentum mode temporarily.

I used momersion with 60 days lookback, 250 days is too much I think.

But I guess you could do a more thorough study.

JayJayD

Pro ,

Hey @Christoffer

Thanks you very much for sharing your experience. I'm very glad you found the Indicator useful.

What I'm testing right now is applying Momersion to a very light smoothed price. In the first example, Harris applies a moving average to a Momersion(250) indicator. I first apply an EMA(10) to the prices and then the Momersion(60) using a 50 as trigger for the signal. I'll tell you as soon I have the results.

JayJayD

Pro ,

@Christoffer,

If the price series is smoothed before the Momersion, it will make the indicator almost always be above 50 because is estimated from first differences multiplications!

Lesson learned: Don't smooth the price series before the Momersion.

Christoffer Sí¸rensen

1.1k Pro ,

Ok.

I guess you have to learn to use momersion like any other indicator.

Just blindly applying it can lead to bad results.

Omar Andres Hernandez

11 Pro ,

Thanks amazing job

JayJayD INVESTOR

Update Backtest

Notebook

To unlock posting to the community forums please complete at least 30% of Boot Camp.
You can continue your Boot Camp training progress from the terminal. We hope to see you in the community soon!

Platform

QuantConnect Community

Alpha League Competition: $1,000 Weekly Prize Pool

Filter Discussion by Tags

Community Ambassadors

Earn free QuantConnect cloud credits by helping others, each week we airdrop credits to the most active community members.

232,274 Quants.

Online Now

Stay Connected