Index | Recent Threads | Who's Online | Search

New Thread
This topic has been viewed 287 times and has 6 replies
mgerstein
Defund the Value Police

Here's a blog on value that might interest you, and more particularly, an explanation of the relationship between value and growth. If you aren't doing both, then you're doing neither.

https://actiquant.com/2020/09/14/defund-the-v...-until-they-get-it-right/

Marc Gerstein
Director of Research, Chaikin Analytics
Blogs: https://actiquant.com, https://portfoliowise.com/portfoliowise-blog/ , https://www.chaikinanalytics.com/blog/
Twitter: @MHGerstein
I predict the future, as soon as it becomes the past

Sep 15, 2020 9:09:01 AM       
Jrinne
Re: Defund the Value Police

Here's a blog on value that might interest you, and more particularly, an explanation of the relationship between value and growth. If you aren't doing both, then you're doing neither.

https://actiquant.com/2020/09/14/defund-the-v...-until-they-get-it-right/


Thank you Marc. Good article!

From the article:
Through this framework, we combine fundamental data with a wisdom-of-crowd approach.

I wonder if—without giving away any proprietary methods—you could expand on how you look at the "wisdom of the crowds."

Momentum? Other technical factors? Consensus estimates? Other data? Maybe none of the above?

And are you always looking for when the crowds are right (and following the crowd’s wisdom) or sometimes investing when you think the crowds might have it wrong?

Probably this next question does get into proprietary methods. But if not (or speaking generally), how does Chaikin Analytics weigh the factors you discuss in the paper? Are they linear rank weightings, like with the P123 ranking system, or do you use a different method?

Thank you in advance for any additional information you can share

Best,

-Jim

From time to time you will encounter Luddites, who are beyond redemption.
--de Prado, Marcos López on the topic of machine learning for financial applications

Sep 15, 2020 10:01:44 AM       
Edit 9 times, last edit by Jrinne at Sep 15, 2020 10:27:33 AM
mgerstein
Re: Defund the Value Police


From the article:
Through this framework, we combine fundamental data with a wisdom-of-crowd approach.

I wonder if—without giving away any proprietary methods—you could expand on how you look at the "wisdom of the crowds."

Momentum? Other technical factors? Consensus estimates? Other data? Maybe none of the above?


All of the above, and it could be anything else anyone comes up with. We don't have a monopoly on the exact factors. The key is to play detective. What factors are telling you about stock behavior consistent with a market belief to the effect that . . . .

And are you always looking for when the crowds are right (and following the crowd’s wisdom) or sometimes investing when you think the crowds might have it wrong?


This is an interesting question. The contrarian game is irresistibly wonderful to play, especially when one is new to investing. So naturally, when I started, at a time when my hair had no grey and when my knees were a helluva lot better than they are now, I had not doubt that the market was populated by an ignorant herd (it helped that there were lots of gurus happy to stoke my ego) and that I was a superior genius. But the longer I've been in tis business, the more I see how Mr. Market, although not always right, is damn smart and entitled to a presumption of innocence. (And it does make sense -- back when all the gurus decided the market was dumb, information was scarce and expensive, so anyone who actually looked at objective info had a bona fide edge. Today, though, that sort of approach is as exciting as black-and-white TV and rotary dial phones.

So we are trend followers (when we use oscillating indicators, it's mainly to look for oversold/overbought condition and identify traing points).

This doesn't make anything easy. There is no actual person named Mr. Market announcing "This is what I believe." So it takes a lot to determine what we should look at to measure the crowd's wisdom, detect when the crowd is losing conviction and considering changing its views, etc.




Probably this next question does get into proprietary methods. But if not (or speaking generally), how does Chaikin Analytics weigh the factors you discuss in the paper? Are they linear rank weightings, like with the P123 ranking system, or do you use a different method?



Yeah, that's proprietary.

Marc Gerstein
Director of Research, Chaikin Analytics
Blogs: https://actiquant.com, https://portfoliowise.com/portfoliowise-blog/ , https://www.chaikinanalytics.com/blog/
Twitter: @MHGerstein
I predict the future, as soon as it becomes the past

Sep 15, 2020 1:54:54 PM       
Jrinne
Re: Defund the Value Police

Thank you Marc. -Jim

From time to time you will encounter Luddites, who are beyond redemption.
--de Prado, Marcos López on the topic of machine learning for financial applications

Sep 15, 2020 2:23:47 PM       
judgetrade
Re: Defund the Value Police

Good article!

Sep 19, 2020 2:24:57 AM       
piard2
Re: Defund the Value Police

Thank you Marc. Very interesting post where statistics meet common sense.

Sep 19, 2020 6:47:18 AM       
Jrinne
Re: Defund the Value Police

Marc,

I hope this is thought of as just good information (or as a shameless promotion for your service if you prefer).

But you can actually download the factors that Chaikin Analytics uses from their site: https://www.chaikinanalytics.com/wp-content/u...-Important-Factors-CA.pdf

This link works for me but I have already given them my email address. You may have to find this on the site.

And I have gotten a personal email message from Mark Chaikin;-)

Anyway good stuff, IMHO.

Very interesting post where statistics meet common sense.

Has anyone ever done factor analysis?

Or noticed that Marc Gerstein and perhaps Marc Chaikin (based on the download) like to group factor into nodes (latent factors)? Anyone who has done factor analysis or principle component analysis knows this can reduce overfitting.

However, Chaikin Analytics use a proprietary method and they may (or may not) have the equivalent of a node. So it may be a bit of a stretch to speculate that Chaikin Analytics benefits from this.

Intentionally or not there is a lot of good statistics (the common sense is a given) going on behind the scenes with Marc’s methods.

Really good stuff.

Edit: Not too shabby of a five bucket (not necessarily an equal number of stocks in each bucket here) rank performance test (below).

Question. If I subscribe to this service can I get an ordering of stocks (best to worse)? and can I get all of the "very bullish" stocks downloaded to be put in as an InList into P123? Or do I only get a readout of each single stock that I enter? That is not clear from the site.

Best,

Jim

Attachment Five Buckets.png (139724 bytes) (Download count: 35)


From time to time you will encounter Luddites, who are beyond redemption.
--de Prado, Marcos López on the topic of machine learning for financial applications

Sep 20, 2020 7:39:27 AM       
Edit 18 times, last edit by Jrinne at Sep 20, 2020 9:48:07 AM