Your model returns stocks you don't like

Hi all,

Here’s the question: You build a model and made your best judgement in it’s constructions based on the backtest results and factor interactions. In the abstract you honestly feel like it’s picking performing well.

But then when you run it “for real” you might see a list of stocks, but maybe 1 or 2 out of 10 would be stocks that make you wonder “why is my model picking this dog?” And you look at the financials and wonder how did my model pick this? You look at the factor ranks but still don’t get it. You see a news story that the CEO even just resigned, sales are flat, and the stock price is about the same as it was 5 yrs ago. So you go in and start tweaking, trying to make sure your model picks stocks closer to your conception of what a good investment would be (by doing things that might disadvantage said “dog”). But no matter what you do, adjustments you make only reduce backtest results. By trying to exclude the “dogs” you end up hurting the performance of the model.

Do you conclude that the original model is better and just take what the model recommends? Or do you prefer the adjusted model with worse returns that tends to excludes companies you don’t like? I guess to what degree do you let these types of reactions to stock selection direct your modeling and investing process? Thoughts appreciated.

For me, I admit to finding it hard to hold my nose and buy a stock that I simply don’t understand - but I also admit that I might could learn to like them if I understood them better. I also see the case that having the stock selections be more in tune with my intuition of the types of companies I’d like to invest in is also desirable in many ways.

Thoughts appreciated.

I check every new stock that shows up as a “buy”. If I don’t like it for what I consider to be a legitimate reason, I exclude it from consideration by using Ticker!=ABC in the universe. A legitimate reason, for example, is that the company is misclassified into the wrong industry. Another example would be that the company has a huge number of warrants outstanding. I exclude companies and even whole industries that I consider to be “commodity” companies. The examples you use like the CEO resigning, flat sales or flat stock price don’t factor into my model so I don’t pay any attention to them.

I used to do this but was proven wrong as often as I was right. I gave it up and now just do what I’m told.

I usually hold my nose too. A good example is LFVN, which has gone up 14% since I bought it three or four weeks ago. I despise everything about this company, which sells snake oil through franchises, ripping off both their customers and the franchise owners. But they were undervalued, so I bought their stock, and I’ve made money off of it. I have to remind myself that LFVN is not seeing my money. Buying their stock on the secondary market isn’t going to affect them in the least. And I’m not going to hold it very long.

But the best thing about finding awful companies in your system is when you figure out how to simultaneously exclude them and improve your system thereby. For example, I wouldn’t have discovered that my system was allowing me to buy companies that were openly talking about bankruptcy until it actually happened. I then put in safeguards against companies with similar qualities (extremely low price, high short-term debt, low Z-score), and my backtests went up.

I think some of us look at risk a little differently than the average retail investor. I also think that many professionals look at risk in the correct way but they have to keep the volatility down to keep investors.

This is Prospect Theory. The general theory regarding how people handle risk won a Nobel Prize (Daniel Kahneman). Studies of how professional traders handle risk and expected return correctly–while retail investor do not—can be found.

I think some risky stocks are underinvested—and therefore undervalued–because investors cannot stand the possibility of a large decline psychologically. This keeps the price low. And the price is below where it should be considering the “expected return.” Sure there can be big declines with these kinds of stocks but the huge pops outweigh the declines if your portfolio is diverse enough and you can stomach the stocks that do not do well.

I am thinking, for example, of a small drug company waiting for FDA approval of a drug. Another example, from may experience was FRO a depressed oil tanker company looking at possible bankruptcy. These are stocks that can have greater than 100% returns with good (or not so bad) news.

I bought ARNA (Arena Pharmaceuticals) before FDA approval of a weight loss drug (before P123 in 2012). I actually did not think they would get approval but when I figured the chances of approval and the prices one could expect with approval (and lack of approval) it was undervalued relative to the expected return. This was despite heavy investment by BlackRock. They had the same impression: it was not just me who had this impression about expected return. Yet, despite the consensus on the expected returns the stock remained undervalued. The average retail investor could not get themselves to invest in this stock or if they did they underweighted it in their portfolio. So, ARNA remained undervalued.

You do not have to be right about what will happen to each individual stock. Just as I was wrong about Arena Pharmaceuticals (I did not expect that it would get approval). After all ARNA had tried for approval for the same drug previously and it had been denied. A drug similar to this drug had been removed from the market due to unwanted complications: death. The stock just has to be undervalued relative to the expected returns.

This was a unique opportunity for me in that the the results were binary and could be calculated easily. And BlackRock agreed in the case of ARNA. So it was an opportunity because much could be learned about what other people were thinking about this stock and my belief about the expect return seemed to be widely held. Most of the time this effect of risk on the average investor cannot be so easily calculated on a spreadsheet. But I think this Prospect Theory effect is still there with more complicated examples.

I think Yuval has said something similar about risk—or something I took as similar. Of course, he uses his own investing terms. Or maybe he disagrees. And of course, anything that removes the real dogs–those stocks with negative expected return–would be good as he suggests in his post.

So, I guess I would first say be comfortable with the out-of-sample results of you port. Take Yuval’s advice: if you can find some red flags that keep you from buying the worst stocks then use it. If you have done both of these: just buy it! If it were an easy decision then the retail investors would have already bid the price up to its fair value.

-Jim

Jim, as I think about it I think my “issues” with ugly stocks and unfailingly following system rules probably stems from a few places:

  1. Seeing ugly stocks in my system makes me doubt my system, and that’s not a good place imho.
    1a) I agree with Yuval that if I can improve the model by removing some of them I will do so, but that is often easier said than done. It’s harder than expected to filter these cleanly. The model is picking them for a reason, even if the appeal is not apparent to me even knowing the model design.
    1b) The model factors make good intuitive and empirical sense, but that doesn’t mean the model ends up buying companies like what I’m expecting it to buy. It’s a surprising result, and reflects how difficult it may be to internalize so many of the relevant investing factors in stock evaluation.

  2. Historically my best investments have never been the crappy-company-getting-a-little-less-crappy variety - so there’s very little intuitive buy-in from monkey brain. When studying deep value, I can intuitively understand the idea of a basket of cigar butts, but have never liked cigar-butt companies and generally had bad experience with them. Some people intuitively love these types of companies, but I guess I don’t really understand them properly to “get” them.

  3. I heard an interview with a quant manager who said he didn’t know what companies his fund was invested in. He knew what the tickers were, but didn’t know the company names, or what they did. He was isolating himself from outside influence and just following his system. I’m not sure that’s a place I’ll be able to get to, but I’m trying to get closer to that. And I’m thinking about how to remove impediments to getting to that point.
    3a) One option is to trade more position with smaller size. With enough position and smaller size it really is easier to just “buy the list.” Negatives are logistics of so many positions and trading costs.
    3b) The other option (and generally better performing option) is having larger positions in fewer companies. That introduces the desire to know more about individual companies which gets me thinking more about individual positions - and I’m trying to think less on the implementation phase of the process.

  4. Unfortunately sometimes intervention has a good result for me. It’s a bad habit I think, but for example I bailed on a poorly trending Argentina investment just luckily prior to a large price drop and the financial crisis they have going down there. I avoided a buy in Brazil for similar reasons, but a Brazil telecom stock is on my list and I’m watching after it’s had a big drop. I think momentum may play bigger part in the international markets than in present domestic market, and have integrated into the models, but sometimes it difficult to ignore odd behavior in price charts accompanied by news even if my efforts at implementing systematic p123 technical rules have been frustratingly unproductive.
    4a) I’ll add I think in watching price action in my portfolio there might be hints when a stock is behaving differently. Kindof like that fish that’s not swimming in time with the rest of the school, or a musical instrument that’s a bit out of tune. I try to keep my mind open to this, and am consciously trying to be more aware when I notice it. I could be fooling myself and it’s certainly not systematic, but it happens often enough with follow-through that I’m beginning to weigh it a bit more.
    4b) Some purely discretionary investments have been some of my best in recent years. It’s difficult to come up with enough of those opportunities though.
    4c) In total, however, I’m trying to get closer to a systematic approach that I don’t think about too much, and just “execute it”. I’m working on rules presently to score myself on and try to balance all of this.

  5. Sometimes there are weird things in the data. When I filter for stocks at extremes sometimes I see unexpected one time shocks that I haven’t controlled for, or maybe just bad data. Over time these might be weeded out as Yuval suggests.

  6. This is rambling on, but I think something I read by Brett Steenbarger is relevant. He suggests difficulty following a trading plan can stem from several sources. Two he cites are a) boredom/need for stimulation (overtrading) as one (probably not me), and b) lack of confidence in system due to not battle-tested, not enough experience with it yet (may be some of that with me because I’m still building out my systems). But underlying most reasons he cites is an unmet need. I though quite a bit about what he was saying, and wonder if a need for creativity is behind my difficulty adhering strictly to a system. I do view these systems as a creative outlet, and I enjoy reading market news and finding opportunities. In a way it is a creative outlet for me, not dissimilar from a game, and I have been focusing heavily on it over the past year. So if my self appraisal is accurate, he suggests to improve adherence to a system I should find other outlets to satisfy that creative need. So maybe going back to songwriting might actually help my trading? Just a thought :wink:

When I first started using P123 I would brag to my wife that I just made $X. She would, of course, ask: ”On what stock?” Like the investor in the podcast I would not know the full name, or often even the sector of the stock. I look that up before I brag now. And I am prepared to “embellish”” when necessary. “Most of their oil fields are in the gulf. Yea, that’s the ticket.”

I have no great purpose in this. This is just laziness.

It could be for the best. I’m pretty sure my discretionary choices would not improve anything.

-Jim

Michael -

I think it’s important to be comfortable with your system. If it’s giving you a lot of stocks that you’re uncomfortable buying AND it’s not making you rich, you need to do some more work on it. That’s a better approach than being discretionary. Be discretionary about the process, not about the stocks that the process gives you.

Here’s what I do. When I set the amount I’m going to buy of a stock, I use an Excel spreadsheet I made that makes a lot of minor adjustments. Some of those adjustments depend on trading difficulties–if a stock looks like it’s going to have a big bid-ask spread or be very thinly traded, I’ll deduct a few points and buy fewer shares. But I also adjust my buy amount based on other factors, including but not limited to ESG and AGR ratings from GMI and executive compensation numbers from Morningstar. I have no idea if doing so improves or worsens my results, because I have no way of backtesting this stuff. But I’ve read over and over again that companies that reward their executives too much are losers, and that companies that pay attention to accounting and governing risk are winners (GMI uses certain factors that we just don’t have access to). Another thing you can do (which I used to do) is to award a company extra points or deduct points based on reviews on glassdoor.com. It’s extremely time-consuming to go through these, but Glassdoor does offer some interesting metrics, so long as enough people have written about the company (if it gets fewer than 25 reviews, its probably best not to take those into account).

On the other hand (and this is the reason I no longer consult Glassdoor), you have to keep in mind that the reason for your success or failure has to do with your adherence to a system which discovers things about companies that are NOT already reflected in their price. Your edge lies in the numbers, not the looks. If you think a company’s products are lame or their management is crazy or they’re outdated or gimmicky, no doubt your opinion is shared by a large number of other investors, and that’s why the company’s price is low. If your system points to undervalued companies, it’s your job to ignore all the sentiment that justifies those prices, and to trust that you’ve found a diamond in the rough.

Your biggest winners are always going to be the stocks that have been underestimated. Don’t underestimate them yourself. Instead, tweak your system so that you’re more comfortable buying them–perhaps in slightly smaller quantities.

Right. I figure I have no reason to believe I’m any better at subjective judgement like this than anyone else is (in fact I’m sure the opposite is true).

Do any of you listen to or read conference calls, and respond in any way to what you hear/see? (Do you consider them useful, unimportant, (or perhaps worse - unhelpful misleading distractions) in the context of your strategy?) Or more simply outside the context of your process regardless of the possible information contained? thanks,

I came across something pertinent to this discussion today. This is from Drew Dickson, CIO of Albert Bridge Capital.

"So our view is this: if more than one of our friends thinks we’ve lost our marbles over a particular idea, we think there is a chance that we are onto something. We believe stock prices reflect opinion first, and fundamentals second. If it feels awful and nonsensical and embarrassing to own something, then we remind ourselves that it feels awful for the market too; and the share price of that awful, nonsensical, embarrassing thing is more than likely reflecting the consensus view.

“As such, if you’re having trouble sleeping, you probably have a more optimal portfolio than if you’re getting seven hours straight a night. This of course is not to say that if a stock has done well, that it can’t keep doing well; nor should we assume that a long-term loser is necessarily going to recover. Moreover, we should all be happy to own something Mr Market already likes if we think that even he is underestimating the fundamental upside. Yet, speaking broadly, it is the non-consensus perspective that is likely to win out, more times than not; and it is only the non-consensus investor who has a chance to outperform.”

See https://www.albertbridgecapital.com/drew-views/2018/6/4/career-risk-alpha-and-contrarian-investing . . .

Thanks Yuval.

Dickson references Kahneman, Prospect theory and risk aversion in his article: “Kahneman and Tversky: An Analysis of Decision under Risk, Econometrica, Vol. 47, No. 2 (Mar., 1979), pp. 263-291.”

But this is just one type of bias. More generally, I like what you said above:[quote]
Your edge lies in the numbers, not the looks. If you think a company’s products are lame or their management is crazy or they’re outdated or gimmicky, no doubt your opinion is shared by a large number of other investors, and that’s why the company’s price is low…
[/quote]
Focusing on the numbers addresses all types of biases (e.g., worrying too much about the CEO who may not have that much impact on earnings). Kahneman calls this bias WYSIATI (what you see is all there is) in his book “Thinking Fast and Slow.” He gives the example of buying stock in a car company because you like the CEO but not paying attention to the earnings.

-Jim

Thanks for the thoughts. I went ahead and bought the specific stock I was referring to yesterday. LYTS is the company I just looked at and gagged, even though the ranking system seems to like it.

[quote]
I came across something pertinent to this discussion today. This is from Drew Dickson, CIO of Albert Bridge Capital.

"So our view is this: if more than one of our friends thinks we’ve lost our marbles over a particular idea, we think there is a chance that we are onto something. We believe stock prices reflect opinion first, and fundamentals second. If it feels awful and nonsensical and embarrassing to own something, then we remind ourselves that it feels awful for the market too; and the share price of that awful, nonsensical, embarrassing thing is more than likely reflecting the consensus view.

“As such, if you’re having trouble sleeping, you probably have a more optimal portfolio than if you’re getting seven hours straight a night. This of course is not to say that if a stock has done well, that it can’t keep doing well; nor should we assume that a long-term loser is necessarily going to recover. Moreover, we should all be happy to own something Mr Market already likes if we think that even he is underestimating the fundamental upside. Yet, speaking broadly, it is the non-consensus perspective that is likely to win out, more times than not; and it is only the non-consensus investor who has a chance to outperform.”

See https://www.albertbridgecapital.com/drew-views/2018/6/4/career-risk-alpha-and-contrarian-investing . . .
[/quote]I don’t know how successful Drew is at investing, but I did play around with backtests of various strategies that clone hedge funds.

Interestingly, I found that buying the stocks held by the broadest consensus of value investors did best. This system did better for example than buying high conviction unpopular picks by good managers. (Furthermore, in my limited and biased experience, battleground stocks–i.e. the stocks that some big name is shorting while another big name is going long–tend not to be good shorts, no matter how great the conviction of the manager is. Examples include Einhorn’s shorts such as Green Mountain Coffee. Take this for what it’s worth.)

As always, I try to fit my theories to the data and not the other way around. After I satisfied myself that this data was accurate, I turn to rethinking the theories. One puzzle I had with the data was to explain how broadly held stocks have more upside. Isn’t it better to be a contrarian?

To answer this conundrum in a way that fits with the data that I have seen, I have two possible theories: One theory is that (A) the consensus of the top investors is contrarian compared to the market. Another possible theory is that (B) it takes time for the price to go up to fair value.

Back in the days when I was picking individual stocks by reading the 10K’s and stuff, I found much greater success buying stocks when fellow value investors agreed with my thesis than when they disagreed. In fact, I found that if most of the investors who did their homework did not like the stock, then it was not likely to ork even if I was sure that it was a good deal. But when the consensus of people who analyzed the stock agreed with me, then the stock would go up; usually sooner rather than later.

I trade all stocks blindly. 100%!!! Very seldem I get curious and digg deper, If I do not like them, I know that I should trade them even more, because whats emotionally hard to trade is an indicator, that you System picks stuff that works.

Take it from Jim Simons. Founder of Renaissance Technologies – the most successful hedge fund ever.

“At the end of a 10-year run, it was clear to me that this gut wrenching business of fundamental trading . . . you know, if you are doing fundamental trading one morning you come in and you feel like a genius. Your positions are all your way. “God I’m really smart. Look at all the money I made overnight.” Then the next day you come in and they’ve gone against you, and you feel like an idiot. We were pretty good at it, but it just didn’t seem to be a way to live your life.

So by 1988, I decided it was going to be 100% models. And it has been ever since. Some investing firms say “Oh we have models” but what they typically mean is that we have a model which advises the trader what to do. If he likes the advice, he’ll take it, and if he doesn’t like the advice he won’t take it. Well that’s not science. You can’t simulate how you were feeling when you got out of bed 13 years ago when looking at historical simulations. Did you like what the model said or didn’t you like what the model said? It’s a hard thing to backtest.

So if you are going to trade using models, you just slavishly use the models. You do whatever the hell it says no matter how smart or dumb you think it is at that moment. And that turned out to be a wonderful decision. So we built a business 100% based on using computer models, starting with currencies and financial instruments, gradually moving into stocks and finally into anything liquid that moved.”

See this video starting about 29:50:

https://www.youtube.com/watch?v=SVdTF4_QrTM

This is a fascinating topic and there is no single correct answer.

At first, pre-p123, I’d individually research and make decisions on each stock produced by a model. The idea was that the model is an idea generator, not a full-fledged investing strategy and that its main function is to narrow the list of stocks worth examining to a manageable number.

Then, my thinking started to evolve as I (still pre-p123) got into manual backtesting. Seeing what a well constructed model could do, and discovering folioinvesting which made it feasible to invest in a full list without breaking the bank on commissions, I started to edge in that direction. I still looked at each stock, but not, the presumption of innocence was much stronger. It took more perceived negativity for me to over-rule the model, but I continued to do it. As it it turned out, though, I did not help my performance and arguably, the pure list was a bit better than my edited list.

Then, as we all know, came p123 and the more serious capacity for fully-automated investing, which i’ve done with real money and with much happiness at the results. Back was doing the Forbes Low Price Stock Report, I felt it was necessary for me to step back to model-and-edit and I did that over five years. But again, tallying up things in retrospect, I could have stayed with the model alone.

Not sure how many folks have seen the new blog site I set up (with a view toward eventual integration into p123) at p123blog.com, but a day or so ago, I published an idea-generator screen addressing small stocks (mktcap<=500, close(0)<10 and acceptable if not always magnificent liquidity). You can check out the article, which provides lots of backtest data:

https://p123blog.com/2018/06/18/understanding-micro-nano-cap-stocks-and-finding-potentially-good-ones/

Now, as I look through the stock list with a view to doing some individual-stock write-ups, I find myself contemplating the exact issue posed in this thread. But then, as I go back and wonder how it happened, I see that I’m getting exactly what the screen asked for and I’m reminded at how badly performance suffered when, during the testing phase, I had rules set in ways that absolutely positively would have populated the list with so-called “better” stocks and “better” companies.

Back to my mainstream; I do as Andreas does, 100% all in. I once wrote an article “Confessions of a Wall Street veteran” in which I explained how I own lots of stocks and can;t tell you, off hand, the name or ticker of a single one, I download Excel from p123 and upload Excel to Folioinvesting.

Before I went into p123 today, I had been contemplating that my new blog post would, instead or a stock writeup, be a post discussing the idea of seeing stinky stocks in a model as a lead in to the individual stock writeups. Well, having see this thread and seeing how it’s on the minds of others, my decision is made. That will be the next blog.

Interesting article, Marc! Thanks for posting.

When I look at the criteria, NOOTC stocks < $10 and $500M mkt cap do OK as a baseline group. But when you add that $150K turnover requirement, the “target list” really underperforms the benchmark. Kudos to you for figuring out how to build a great screen from a poor performing target list with the use of ranking systems.

When you post your screen backtest results as in the blog post, what price are you assuming? Next open? And do you alter the slippage and carry cost defaults?

Many thanks.

[quote]
At first, pre-p123, I’d individually research and make decisions on each stock produced by a model. The idea was that the model is an idea generator, not a full-fledged investing strategy and that its main function is to narrow the list of stocks worth examining to a manageable number.
[/quote] My actions were similar, as were probably a lot of others’. Now I seldom question the recommendations, not because I have necessarily proven them to be consistently superior but because it is so much easier and results are usually good.

With the flourishing of robo advisor portfolios that are heavily quant driven to minimise management expenses, perhaps the underlying market support for stinky but statistically promising stocks is becoming stronger? I doubt that it hurts!

But there’s a behavioral issue involved as well. Take this scenario. You’ve lived all your life never getting paid more than $100K a year. You’ve saved diligently, and now you have $700K to invest in stocks. Your best model has you putting 6% to 7% into each position, and it’s a model that really works, that consistently beats the market big time in real time. So now you’re putting half a year’s salary into a hold-your-nose stock. You tell yourself it’s all about the process, and you’re doing fine. And then someone asks you, so, what companies are you investing in these days? XXX is your #1 stock? Are you kidding? Well, they say, as long as it’s only a small position, I suppose it can’t hurt too bad. What? You’ve put $45,000 into an unscrupulous company that rips off its franchises and sells snake oil? Are you nuts? Well, you say, I also have a lot invested in YYY, who bills their customers for thousands of dollars for a product that should cost $50 and then refuses to return their calls. And their CEO is a shady guy who once fired an employee because she was black. OK, well, that’s only two companies. Oh, but then there’s company ZZZ, who basically deal in stealth ads, the kind that get into your computer and pop up when you don’t want them to. And then there’s the mining company that mistreats its workers. And so on . . .

I’m not about to stop putting huge sums of money into stinky stocks, but I have to have a very dark sense of humor about it.