How to Test My Thesis About Betting Against Twitchy Markets

December 6, 2016 7:52 AM

Thesis: highly successful public companies trade on a hair-trigger. This ensures overreaction to immaterial or near-immaterial negative news, as twitchy traders on both sides rush to anticipate short term reaction (betting on the betting rather than on the bet). Investors with a more patient attitude who buy these dips should, in the aggregate, profit when the stock recovers. So...is there a modeling service I can use to test this thesis?

No morsel of news, no matter how trivial, seems to be ignored in the moment-to-moment stock price trajectory of major companies (particularly ones with high share prices, where everyone's extra-nervous and twitchy). I suspect this offers an advantage to more patient investors.

Taxes and transaction costs would be a concern. But the real pitfall would be getting caught "holding the bag" if stock prices continue to sink rather than recover due to 1. overall market movement, 2. other/unrelated company news, 3. insider trading, or 4. buy/sell timing (it would take time to ascertain the immateriality of a given reaction, and price would be a moving target as I react and attempt to trade. And that's just the buy side of the investment. How long would I want to await recovery before selling?).

Factors 1-3 could also work in my favor, and factor 4 perhaps could be mitigated via algorithm (I'm not sure how I'd create or implement that, though...it's over my head)

But if my thesis is correct, and stock prices twitching upon immaterial negative news recover more often than not, there's got to be a way to profit. So I'd like to test the thesis somehow, with a reasonably broad sample set.

On the odd chance you're a professional curious enough to research this yourself, please PM me the upshot! I promise I won't personally move the market! :)

posted by Quisp Lover to Work & Money (18 answers total)

Have you read the research on this? There is a lot of it.
posted by JPD at 7:56 AM on December 6, 2016 [4 favorites]

IIRC, there's various "Virtual" or "Fantasy" stock market games that use real stock data. Maybe play with one of those for a while?
posted by SansPoint at 8:01 AM on December 6, 2016 [2 favorites]

Response by poster: JPD, I've read nothing. I'm just going on my own observation and intuition. If you could point me to something relatively non-technical to read (or synopsis), I'd be appreciative.
posted by Quisp Lover at 8:04 AM on December 6, 2016

Best answer: Google short term mean reversion in share prices.

Basically. Yes in some form this exists but that's a vast oversimplification of it, and wall street is rife with people trying to do it in very sophisticated ways.

Google Stat Arb as well.
posted by JPD at 8:38 AM on December 6, 2016 [3 favorites]

This is not a unique insight. It's one category of market timing. You can make boatloads of money doing it, but you have to find a way to do it better than all the other hedge fund managers, high speed trading syndicates, math PhDs, and insiders who are attempting to do the same thing.
posted by Winnie the Proust at 8:39 AM on December 6, 2016 [2 favorites]

Best answer: In answer to your question: you can test a trading thesis on Quantopian. Yours is underspecified though, so I'm not sure where you'd begin.
posted by caek at 8:42 AM on December 6, 2016 [3 favorites]

You could test your idea fairly easily using Quantopian, a free site.

I think you'll find that it's not correct though. If it were that easy, a few people would be running this algorithm, which would in turn push prices up at signs of bad news and make it not work any more.
posted by miyabo at 8:42 AM on December 6, 2016

Response by poster: ========
You can make boatloads of money doing it, but you have to find a way to do it better than all the other hedge fund managers, high speed trading syndicates, math PhDs, and insiders who are attempting to do the same thing.
========


Thanks. But I'd counter-argue that most of those hedge fund managers, high speed trading syndicates, math PhDs, and insiders are on the other side of this equation: they're the twitchy short-term parties betting on the bets. There may be patient traders on my side of the trading, but they're not the ones moving the stock price...hence the thesis (original or not).

If the preponderance of trading is consistently irrational, there's profit to be made, period. And that preponderance is gauged exclusively and entirely on price movement. So the real question is: am I correct that this is irrational (beyond the short term with which a disproportionate number of traders are preoccupied)?

========
a few people would be running this algorithm, which would in turn push prices up at signs of bad news and make it not work any more.
========

The ultra short term trading is immense on established stocks. I'd imagine it would take way more than a few people nibbling at this thesis to meaningfully offset their impact.
posted by Quisp Lover at 8:46 AM on December 6, 2016

By all means learn about this stuff! It's interesting! It affects us all. But to cut to the chase: you're going to learn that your idea is "do what fabulously well-funded and experienced people are already doing... but better".
posted by caek at 8:51 AM on December 6, 2016 [14 favorites]

Stocks also pop upwards on short term good news.

Your everyday long term middle class investor already does what you are suggesting, or their mutual funds do, anyway, by dollar-cost averaging investments over time, with minimal reactivity. Short term churning is always gambling (and also racks up trading fees) The folks with better math skills are always ahead of you.

The only way to guarantee a quick small fortune on Wall Street remains to start with a large one and a hunch.
posted by spitbull at 8:54 AM on December 6, 2016 [2 favorites]

But I'd counter-argue that most of those hedge fund managers, high speed trading syndicates, math PhDs, and insiders are on the other side of this equation: they're the twitchy short-term parties betting on the bets.

You can't trade fast enough to beat the professionals who have already had this insight and are attempting to execute the same strategy. Keep in mind that these "twitches" could last for seconds or less. The professionals will have computers tracking this in real time and, when they see an opportunity, can execute a trade at absurd speeds. You might benefit from reading Michael Lewis's Flash Boys, though you should take it with a grain of salt.

Keep in mind that even with those technical advantages most hedge funds still don't outperform stock indexes over the long haul after fees! There is a strong survivor bias in reports on the industry.
posted by praemunire at 9:07 AM on December 6, 2016

Mod note: One comment deleted. AskMe's not a discussion or debate space, so OP please don't get into a back-and-forth over the merits of your thesis, and commenters, please let the "this is a bad idea" thing stand as stated and from here on, focus on things OP can read and/or where to test.
posted by LobsterMitten (staff) at 9:23 AM on December 6, 2016 [1 favorite]

I also recommend Michael Lewis's Flash Boys, it really goes into the current state of the art of making money on twitchy stocks.
posted by sideshow at 11:07 AM on December 6, 2016

I've wondered about similar strategies. As noted above, there are already multiple traders trying to do most anything you can think up, so I wouldn't be optimistic myself.

I read a book on commodities trading that made a strong argument the the only way to succeed is to have information that is not yet discounted by the market. The commodities company I was working at back then made a proprietary comprehensive forecast of the grain markets, and I took that as confirmation the information is the key, not timing.
posted by SemiSalt at 11:48 AM on December 6, 2016 [1 favorite]

I'd suggest reading The myth of the rational market which is a sort of intellectual history of trading theories. You can get a survey of what other people have tried and what the bar is you need to clear to make a profit, even off of accurate and true intuitions.
posted by mark k at 12:35 PM on December 6, 2016 [2 favorites]

I've used this strategy professionally. I would use one of the above-recommended links. I also recommend researching specific events to try to uncover every single factor in the moves. Your theory is wrong but that doesn't mean you can't make money by buying when the price of an individual stock drops.
posted by michaelh at 1:52 PM on December 6, 2016

Pricing happens on all sides of the equation. Bid prices, ask prices, selling, and not-selling (and more) all affect the price beyond reactions to information.

Keep in mind that traders will go as far as selecting office space near the exchange so that the network distance from them to the trade is shorter than other peoples'. You will have to beat these guys, even if your algo is better, because a worse algo that acts quicker will confound your model when the price changes under it.
posted by rhizome at 5:07 PM on December 6, 2016 [1 favorite]

Also came in to recommend "Flash Boys" and that you spend some time reading about high frequency trading and latency arbitrage. What you are describing - anticipating a reversion to the mean - was a viable strategy for human traders in the 1980s. In a gradual arms race of computing power, the twitchy traders you describe are now actually twitchy computer programs running on servers in the basement of the NASDAQ where the lead time for a news event is now measured in milliseconds, rather than seconds or minutes. Also, you - as an individual - would have a higher transaction cost than established players, and that really adds up in any kind of HFT scheme. Bottom line: rather than trying to simulate your thesis, I'm suggesting that you read some time learning what others have tried and how trading has evolved with technology. Finally, I'd also suggest reading "Reminiscences of a Stock Operator" by Edwin Lefevre. Originally written in 1923, it is amazing how much has *not* changed about markets and trading in the last 100+ years.
posted by kovacs at 7:16 PM on December 6, 2016 [1 favorite]