investing on losers

October 27, 2008 1:44 PM

Investing/Day Trading -Why is it bad to pick stocks off the losers of the day assuming they will rebound?

I have been casually watching stocks on Google's daily losers list, and playing with $1000 of pretend money and Excel.. and so far I am up 50% on three weeks (not counting $9/trade fees).

So -- Why doesn't this system work? Bonus points for any investing research that shows its a bad idea (tm).

Naturally there are some mega losers.. but even back when Fannie May tanked.. there would have been money to be made by buying at the crest of the loss, and selling the next day.

For example... If I had invested $1000 in Fannie May at the end of trading on September 8th, and sold it at the end of the day on September 9th I would have netted $356 (minus trade execution fees). I caught Foundry Networks as they started to go up a second time on the 24th of October @ $10.85 and now they are at $12.39.

Naturally there are really bad examlpes of this (like the drug company Savient, but even that might go up tomorrow! It seems like people overreact on large percentage sales.. and then it rebounds.

Am I playing with fire that will burn me? Is this how day trader junkies get hooked?

posted by SirStan to technology (29 comments total)

Because your gains will be substantially wiped out by the capital gains tax.*

*This is a feature, not a bug.
posted by stet at 2:00 PM on October 27, 2008 [2 favorites has favorites]

Read up on contrarian investing for more info.
posted by TedW at 2:06 PM on October 27, 2008

Generally, timing the market on short-term volatility is a losing game. Philip Feigin said it best, imho:

"For the typical retail investor, day trading isn't investing, it's gambling. If you want to gamble, go to Las Vegas; the food is better."

It sounds like, to me, you're trading based only on the loss or gain of shares, and know next to nothing about their fundamentals, business practices, revenue/profit/loss projections, position in the market place, etc. etc. Even trained day traders, who know how to play the market (of which I am not one, and don't know how to play), are getting their asses handed to them right now, in many, many cases.

You're not playing with fire as long as you view the money you're investing as you would money taken to Vegas -- if you can't afford to lose it all, don't bet it.

And, to answer your original question, because they don't always rebound. I've watched Sirius/XM drop, and drop and drop, waiting for it to bottom -- which it finally did on Friday. However, in the mean time, it's lost 70% of its value in the last five weeks.

*I am not an expert by any means AT ALL, but have dabbled for eight years in long positions that have made me decent returns. When I've made snap decisions, it has invariably been a poor decision (oh, those shares of Apple at $46 pre-split.....sigh.....)
posted by liquado at 2:08 PM on October 27, 2008 [1 favorite has favorites]

That's Warren Buffett's advice, anyway: Be fearful when others are greedy and greedy when others are fearful."

The problem is figuring out which stocks tanked as part of a larger sell-off or because they're undervalued, and which stocks tanked because they were inflated to begin with. But the philosophy of "buy low, sell high" works pretty well, as long as you can figure out where low and high are.
posted by Airhen at 2:15 PM on October 27, 2008

I wouldn't say that buying stocks when they had a bad drop, then selling then *the next day* is something that Warren Buffett would advise.
posted by thewalledcity at 2:18 PM on October 27, 2008

Just ask yourself: "What are the chances that I, SirStan, unlike 10 million other people playing the market, have figured out a way to consistently beat the market?"

I don't think it's likely.
posted by mpls2 at 2:18 PM on October 27, 2008 [2 favorites has favorites]

Why is it bad to pick stocks off the losers of the day assuming they will rebound?
Quite simply, it's because you have no way of knowing whether the stocks are really going to rebound or just keep falling.
posted by peacheater at 2:32 PM on October 27, 2008 [2 favorites has favorites]

I wouldn't say that buying stocks when they had a bad drop, then selling then *the next day* is something that Warren Buffett would advise.

Ah, sorry, I was referring only to the general idea of picking the losers.
posted by Airhen at 2:33 PM on October 27, 2008

I have a policy of doing exactly this, when I am absolutely certain it will come back the next day. I've done it twice in two years. I've got a .500 batting average.

Each time I put a significant amount of research into my purchase. I don't expect that you have the time to do the research every day.
posted by Pants! at 2:41 PM on October 27, 2008

If you want the orthodox economics answer, it's a bad idea because of the efficient markets hypothesis, which says that the equilibrium market price will reflect all information that is known about the "true" value of the stock by any and all participants in the market. One major implication of this is that over time stock prices are expected to follow a random walk. In other words, one period's value will be the previous period's, plus or minus completely random noise. Short-run movements in one direction or another are therefore entirely meaningless.
posted by shadow vector at 2:43 PM on October 27, 2008 [1 favorite has favorites]

The terms you want to use in your search are "serial correlation" and "serial autocorrelation." I suggest heading to a library with access to academic research.

There has been a lot of scholarly work performed on the subject. The conclusion, IIRC, was that the effect used to be very large and has gotten much smaller over time. Remember that if an effect like this reliably exists, it will be exploited in the market by the party(ies) with the lowest transaction cost. Which is not you.
posted by Kwantsar at 2:46 PM on October 27, 2008 [2 favorites has favorites]

Summarized:

- capital gains tax reduces your gains, does not reduce your loss
- transaction costs reduce your gain and your loss
- you are unlikely to reliably pick the stocks that lost today, but will gain tomorrow

and

- your bank (cash reserves) are lower than the rest of the market, meaning that even if you are right in the long run, in the short run you will go bankrupt and be unable to participate in the long run outcome.
posted by zippy at 3:17 PM on October 27, 2008

Messed up on transaction costs, meant to say:

- transaction costs reduce your gain and your increase your loss
posted by zippy at 3:17 PM on October 27, 2008

Seconding zippy: "your bank (cash reserves) are lower than the rest of the market, meaning that even if you are right in the long run, in the short run you will go bankrupt and be unable to participate in the long run outcome." I mean, in addition to all the other points. :-)



Incedentaly, this is not just a problem for small-time investors. Everybody's ability to predict what will happen "in the long run" can be deeply flawed. I read recently that in this financial crisis "the banking system seems to have lost more on risk taking (from the failures of quantitative risk management) than every penny banks ever earned taking risks". It doesn't matter if you think you would win on average: a rare event can be so very bad that it will dominate your results. Your system can work great until it doesn't, and then you're bankrupt, and you may not even be able to calculate the liklihood of that happening.

(See the link for a much better explanation of this concept. Another example from the article: a turkey can spend 364 days learning that the world is a good, safe place, that she'll wake up each morning and get fed. She simply has no means of factoring into her calculations the probability of what will befall her on the fourth Wednesday in November...)
posted by wyzewoman at 3:38 PM on October 27, 2008

This is normally called a liquidity providing strategy. See Khandani and Lo, 2007 (PDF) provide an examination of it starting on page 7 ("Anatomy of a Long/Short Equity Strategy") - "The high turnover and the large number of stocks involved also highlight the importance that technology plays in strategies like (1), and why funds that employ such strategies are exclusively quantitative."
posted by milkrate at 3:39 PM on October 27, 2008

To all those saying "Capital gains tax" is the reason this won't work - why is that, exactly?

Can you not write off capital loss against capital gains?

(If I lose $100 on stock A and gain $100 on stock B, am I not coming out wiht a net captial gain of 0?)
posted by TravellingDen at 4:07 PM on October 27, 2008

If you enjoy this sort of thing, try handicapping horses. It costs $2.00 to start, the ADW operations will pay you to play (around 5% of your investment, depending on the track) the information is readily available from a variety of sources for you to make informed investments, and it only takes a couple of minutes to see how you've done.
posted by Floydd at 4:07 PM on October 27, 2008

If I lose $100 on stock A and gain $100 on stock B, am I not coming out with a net capital gain of 0?

Good luck on that strategy. 1000 times zero is still zero.

Assuming you actually do make a profit instead of breaking even, your net gains are taxed at your marginal tax rate, say 25%, leaving you with only 75%. This is because, as a day trader, your gains are taxed as ordinary income, instead of long term gains at the 15% rate.
posted by JackFlash at 4:41 PM on October 27, 2008

If you want the orthodox economics answer, it's a bad idea because of the efficient markets hypothesis, which says that the equilibrium market price will reflect all information that is known about the "true" value of the stock by any and all participants in the market.

Of course, the idea that there exists such a thing as an equilibrium market price, as opposed to a random walk based on an unpredictable mixture of research, fear, greed and herd instinct, is pure assumption.
posted by flabdablet at 5:57 PM on October 27, 2008

So you make less than you would in an imaginary world with no taxes. Why is that an argument against it?
posted by smackfu at 5:57 PM on October 27, 2008

wyzewoman points out a good author. If you want your mind blown, read The Black Swan.
posted by wastelands at 6:01 PM on October 27, 2008

There are lots of theoretical reasons why this shouldn't work, but if you enjoy thinking about these sorts of things, rather than paper trading with Excel for a few weeks, you might enjoy downloading a program like Metastock or eSignal. These will let you write up a rule or set of rules like yours in a simple scripting language, and then test the returns that strategy would generate over actual market data over a long time. They are also priced for non-pro's to buy, but I don't remember the actual numbers. You can do your own research!

The short answers to your questions are "yes, you are playing with fire that will burn you" and "yes, this is how day traders get hooked."

The strategy you are talking about is a very simple example of what is generally called, depending on where you take it from here, "technical analysis" or "statistical arbitrage". The financial markets people have never been much good at naming things, as TA is anything but technical and stat arb is anything but arbitrage (ask anyone who was invested in a stat arb fund late summer '07 for proof). One thing you'll learn if you google up those terms, particularly stat arb, is that there are many simple stat arb strategies that are profitable for small amounts of money over some period of time, but then become unprofitable. Unfortunately, the market doesn't necessarily go out of its way to tell you when that is going to happen. The old joke about this is that most stat arb guys are long volatility, but short volatility of volatility. Stat arb guys call this "regime switching". It is interesting that you started to look at this particular phenomenon in this market, because this is the type of market where this strategy is likely to succeed for a while.
posted by jeb at 7:36 PM on October 27, 2008 [1 favorite has favorites]

The real question is does it perform any better than radom stock selection, or picking the winners etc. To follow up on jeb's line, take the factor risks out.

I can definitively tell you that (in aggregate) this strategy does not make money but cannot provide the research.
posted by H. Roark at 8:33 PM on October 27, 2008

Thanks for the feedback ... some excellent information to go read as well!
posted by SirStan at 8:44 PM on October 27, 2008

@stet - capital gains tax wont be that bad .. and its only on net profits. It should be marginal at this level of playing.

@TedW - Added to read queue. Tahnks!

@liquado - I can't goto vegas on work time! Good points on not playing with this months rent money...

@mpls2 - I doubt it as well.. that doesnt mean I cant play arm chair investor :).

@peacheater - isnt that the hook for investing? :) Its cheaper than scratch tickets, and a better chance of winning!

@Kwantsar - I was looking for research on just that topic! Any specifics would be wonderful.

@milkrate - appreciate the PDF linK!
posted by SirStan at 9:24 PM on October 27, 2008

One question that isn't clear is how you decide when to sell the shares. If you automatically buy in the am and sell in the pm you aren't giving the stocks any time to bounce back if they have good fundamentals. If you sell once you have a gain and hold on to the stock that is down hoping it will turn around, you will soon end up with a portfolio of losers.

Second suggestion is to take your Excel spread sheet and add in trading costs and then subtract capital gains to see what your actual profit was on those three weeks. It will not tell you how good your strategy is in the long term but it will give you a better read on how your hypothetical portfolio performed.
posted by metahawk at 11:36 PM on October 27, 2008

You've had theoretical and orthodox answers, no point in repeating them. The colloquial phrase for what you're trying to do is 'catch a falling knife.'

Think about it for a minute.
posted by ikkyu2 at 11:43 PM on October 27, 2008 [1 favorite has favorites]

@ikkyu2 - somewhat.. but its more like.. catch a bouncing ball.. on the way up... :) Sometimes you might snatch at a ball that isnt there and it keeps falling.
posted by SirStan at 7:02 PM on October 28, 2008

No, you've misunderstood me. I am not trying to bring an old-fashioned saying about knives to describe your idea. I am telling you that stock pickers refer to the precise strategy you're trying to implement as "trying to catch a falling knife." They use this particular metaphor because what happens is you get cut and the knife keeps falling.
posted by ikkyu2 at 9:58 PM on October 28, 2008 [1 favorite has favorites]