Two new stock models

I’m pleased to announce that Portfolio123 has launched two new stock models of my own design.

Portfolio123’s stock models are free models that anyone can invest in. All you have to do is sign up, open an account (real money or paper), and follow the strategy using the “Invest” side of the website (or simply press the “Invest” button on the links below). We’re hoping that these models will attract a lot of new visitors and intrigue them enough that they’ll want to develop their own models.

“Gems of the S&P 1500” (https://www.portfolio123.com/app/investment/add-new?browse=1&t=STOCK_STRATEGY&pid=1567373&sst=1) is based closely on a designer model I launched in May 2017. The only real difference is that I’ve expanded it from a ten-stock to a fifteen-stock portfolio. This high-turnover model invests in safe, high-quality small caps and mid-caps from the S&P 1500, a group of stocks subjected to rigorous standards from the outset. You can read all about it here: https://help.portfolio123.com/hc/en-us/articles/360025483932-Gems-of-the-S-P-1500.

“Small-Cap Quality” (https://www.portfolio123.com/app/investment/add-new?browse=1&t=STOCK_STRATEGY&pid=1567374&sst=1) is a brand-new thirty-stock model that focuses on stocks with market caps between $200 million and $10 billion with high liquidity and low bid-ask spreads. We all feel that the best opportunities for making money in stocks lie in small caps, which are more responsive than large caps to the kind of data we offer. This is a medium-turnover model. You can read all about it here: https://help.portfolio123.com/hc/en-us/articles/360025780451-Small-Cap-Quality.

If you don’t want to be so aggressive, you should check out the other free stock models—we have a total of nine of them, most of them in large caps with low turnovers—as well as eleven different ETF models.

Oh, and a word about the holdings listed in the models. The market cap of any stock less than $1 billion should be multiplied by 100 to get the real market cap. This is a minor bug; it should be fixed in a day or two.

Yuval,

I notice your models are all heavily based in deep fundamentals with little reliance on price or technicals. You also mention that turnover is somewhat high and holding periods are several months. Can you comment how these 2 things fit together, in terms of most fundamentally based systems needing significantly more time for fundamental value to be realized than several months. I would think this would be even more significant in the small / micro cap space where many companies go undiscovered for extended periods of time. Thanks

That’s an excellent question. These models hold some stocks for a long time–Small-Cap Quality has held on to EIGI for over a year and to NRIM, RVSB, and PCOM for over six months, and four out of the fifteen stocks in Gems of the S&P 1500 have been held for over six months. But in general, they’re responsive not just to value ratios but to the indications of growth in quarterly earnings reports and to a wide variety of sentiment indicators.

It’s very important for investors to understand that large caps move in price very slowly and you want to hold onto them for years, while microcaps and small caps move very quickly and it’s best to be in and out. If you’d like to read why high-turnover strategies are more profitable in small-caps in general, I’ve written an article about it, which you can read here or here.

Do stock prices respond to data and or does data respond to stock prices? By responsive to, do we actualy mean predictive of?

Just some thoughts on scoping the discussion.

A great deal of the data that investors rely on isn’t so much “responsive to” stock prices as “based on” stock prices. Value ratios and momentum measures are good examples. It would be awfully hard, however, to say that quality measures respond to stock prices. The price of a stock will have little to no impact on asset turnover or free cash flow. There are middle grounds, too. Per share measures will be affected by stock buybacks, which are responsive to stock prices. Stock buybacks can also have a major impact on the balance sheet, as they lower the amount of cash on hand. So it’s not a completely clear picture.

On the other hand, stock prices are certainly affected by data, especially when that data is seen as predictive of a company’s prospects.

David,

Could be we are taking advantage of this without knowing it.

Of course, we make money only by: signal → price change. The signal has to come first.

But sometimes: signal = price change, or price change → (more) price change—better known as momentum or sometimes reversion to the mean.

But as you allude to it could be price change → factor → price change. A simple example would be a price change causing an analyst to adjust his/her price target leading to further buying from investors following the analyst.

But I suspect it is often factor1 → price change → factor2 → further price change. Factor1 could be any number of factors we use but the initial price change could cause heavy covering of the shorts or a short squeeze (factor2) for example.

But, in truth, I suspect there are many at P123 with their own factor1 and factor2–along with a momentum factor—taking advantage of this and not even knowing it.

One of the strengths of P123. We do not always have to know why something works to make money from the interaction of factors. Heck, we can even have the wrong idea of why something works or think it doesn’t work and still profit.

The magic of machine learning.

-Jim

Yuval - do either of these models use market timing?

thanks, Debbie

No, they don’t.

Is there a way to put these strategies in a book? I would like to look at leverage and correlation to other models.

Walter