At any given time, I will have 45-60 companies that have passed at least one of my screens in the recent past. This is my investible universe. I am comfortable that over time, this group of companies will beat the market.
However, I also know that in any given year, some percentage of my stocks will not beat the market, and a subset of those will actually lose money.
So I am interested in the concept of using P123 ranking systems on a periodic basis (weekly or monthly or quarterly) to manage my portfolio and eliminate the worst stocks from the companies that pass my screens.
By “worst,” I am hoping to forecast the biggest losers over the next time period and eliminate those.
Does anyone have any experience with the P123 rankings in this regard? If so, which ranking systems are best for identifying future poor performers?
I would think that the Basic Momentum rank would be alright, but there’s always that pesky mean reversion risk.
Another way to look at it is to build a short ranking system off your universe of stocks. This is not easy and I have not seen many successful shorting systems especially with Large Caps. I think you would have better luck building a long ranking system that picks the best stocks from your universe and you pick the top 10 or 20. Don’t worry about the biggest losers try and pick the biggest winners.
Forecasting declines is tricky at times when the market has an upward bias because a lot of stocks will rise in line with the market despite the absence of any noteworthy bullish characteristics. But if you can work with the idea of eliminating the most likely underperfromers, I think that can be in interesting direction in which to go. Perhaps you can start with the Basic: Quality ranking system and tweak. That, in conjunction with downward estimate revision and bad high valuation could be interesting. And I would recommend considering earnings quality: this is an area in which eliminating the worst may wind up more productive than picking the best (try the pre-built beneish factors). Look, too, at Piotroski. As to momentum, play mean reversion fro the other side; look for spikes in shares that make them ripe for correction.
You probably should build your own multi-factor “Dump the Dregs” ranking system and, once you establish the universe you want, use a factor like Rank<25. Then, once you have your per-qualified (based on your other screens) mini-universe and eliminated the worst 25% as per your new raking system, then you can use a more conventional buy-oriented ranking system to pick the best X number of stocks for your portfolio.