EDIT: Based on criticism below, I am reconsidering how to proceed with this idea (or something in the neighbourhood).

A topic that has been on my mind recently is where, in our complicated lives, there might be low-hanging fruit ready to be picked by a motivated rationalist. Actual, practical, dollars-and-cents fruit.

In possibly-related news, here is how the writer of About.com's beginner's guide to investing describes the stock market:

Imagine you are partners in a private business with a man named Mr. Market. Each day, he comes to your office or home and offers to buy your interest in the company or sell you his [the choice is yours]. The catch is, Mr. Market is an emotional wreck. At times, he suffers from excessive highs and at others, suicidal lows. When he is on one of his manic highs, his offering price for the business is high as well, because everything in his world at the time is cheery. His outlook for the company is wonderful, so he is only willing to sell you his stake in the company at a premium. At other times, his mood goes south and all he sees is a dismal future for the company. In fact, he is so concerned, he is willing to sell you his part of the company for far less than it is worth. All the while, the underlying value of the company may not have changed - just Mr. Market's mood.

I have heard this narrative many times before, and I'd like to test whether it is accurate - and in particular, whether LWers can consistently beat the market.

The skeptic may well ask: why should LWers have an advantage? Why not go to the professionals - investment advisors? Also, isn't there a whole chapter in Kahneman about how even smart people suck at picking stocks? And what do you, simplicio, know about this anyway?

LWers may have an advantage by virtue of being educated about such topics as cognitive biases, sunk cost fallacy, probabilistic prediction, and expected utility - topics with which investment advisors et al. may or may not be familiar on a gut level. I am not sure if we're any better, but I'd like to test it. Also, if LW turns out to be any good at offering such advice, that advice would presumably be free, unlike that of yon advisor (fees tend to kill returns on investment - just ask anybody who uses Intrade). As for what I personally know - not very much yet. But I find competition very stimulating.

Accordingly, my proposal is for a contest: over the course of 2013, I will set up & maintain a Google Drive spreadsheet. This spreadsheet will be shared with contest participants. Each participant will have say $5,000 of play money to use "buying" (or "selling") stocks on the exchange of their choice. Contestants will record the date of purchase or sale, quantity, and preferably provide comments regarding why they are buying or selling.

At the end of this contest (Dec 31, 2013?), I will commit to Paypal the winner (defined as the person with the highest market valuation of play assets as of midnight on that date) the equivalent of $50 CAD in their local currency. In the unlikely event that I win, I will donate that $50 to the Against Malaria Foundation. (Above commitment does not take effect until I actually gauge interest in this contest, figure out an end date & rules etc., and decide to proceed. If anyone else wants to throw money in the pot, please do.)

The purposes of this post are therefore:

  • to find out who is interested - please leave a comment below, and e-mail me at ispollock [at] gmail.com if you want in;
  • to solicit constructive and destructive criticism of the project, especially from any local experienced investors (in particular, perhaps a one-year timeframe is too short for a meaningful contest? Also, real-world experience of transaction costs in buying and selling would be extremely helpful);
  • to ask if anyone knows of a better software platform for the contest than Google Drive, or knows of any extremely helpful resources I should be reading/linking to.
I am probably just being naive, but I am rather excited about what LW could accomplish here. Even an abject failure would be instructive, if not inspiring.

New to LessWrong?

New Comment
31 comments, sorted by Click to highlight new comments since: Today at 1:46 PM

If more than a few LW'ers participate, the reliably best strategy to win $50 with play money would be to just invest in some very volatile option or other derivative, which'll either win over any conservative strategy or bomb out. Also the least amount of work, fire and forget.

I remember we held a similar contest in 5th grade - everyone paired up and picked some stocks out of the newspaper, and then at the end of a month investing the winner got some candy. The winners just invested 100% in Intel :)

full disclosure: due to earning money in two different currencies (local + remote job) I must engage in forex market to some degree. This lasts for over 2 years now so I think I have some experience now. (Yes, it's a bit different than stock market, but still)

Almost everyone wins with toy accounts, but not with real money. People usually don't understand fully why that is. Let's get the obvious out of the way: on a toy account you think: "I'm going to test this strategy!" With real money you think: "f###, f###, f###! If I don't sell now and it drops more, I'm going to loose 3 months of income!" - that's the usual risk aversion everyone knows about. It's only devastating at the beginning - after a while you'll train it away to acceptable degree.

Now consider not-so-obvious stuff that doesn't happen with toy accounts and never goes away:

  • you crashed your car and you need to sell some stocks at whatever price they're currently at to buy yourself a new one
  • you found a bargain sale (say, a house) that you'll be able to resell for twice as much next year
  • you came up with an idea for a new cool product. If you don't invest in it now, someone else will

In summary - the hidden cost of freezing your assets until your prediction/strategy is profitable (even assuming it's correct) is often higher than you'd like, not to mention time you need to put in evaluating this mess. This is the reason why I currently use forex only for hedging strategies (buy/sell currencies ahead of time at good price - when I know I'll need it in the future). But that's not even possible for stock market.

My advice: forget this thing. Just do what what you love and what you're good at: learn a skill or language, change job, start a business - invest in yourself! If you're not inherently interested in finance then stock market will be a waste of your time - you could use this time better (but not easier!) in your own field. Market offers a false promise of effortless gain.

In summary - the hidden cost of freezing your assets until your prediction/strategy is profitable (even assuming it's correct) is often higher than you'd like, not to mention time you need to put in evaluating this mess.

As a wise guy once said, "The market can stay irrational longer than you can stay solvent."

At the end of this contest (Dec 31, 2013?), I will commit to Paypal the winner (defined as the person with the highest market valuation of play assets as of midnight on that date) the equivalent of $50 CAD in their local currency.

Bad idea. That would reward risk-seeking behavior, which is the opposite of what we usually want.

It's almost irresponsible to discuss this without addressing Efficient Markets, which more or less suggests that this is a bad, no good, terrible, rotten idea.

I only half mean to be rude, and don't mean any personal offense.

You're right.

I've only met a few professional stock-traders, but of the two that I asked, both were aware of the heuristics and bias literature.

That's not surprising.

I'm fairly skeptical whenever someone says "Even though there is no evidence that anyone has ever been able to reliably do , I might be able to do it because I understand ", including for X = "beating the stock market without inside information" and Y = "heuristics and biases literature".

That's not surprising.

It was to me. BTW, by "were aware of", I mean "had heard of".

A major problem with stock picking and other similar forms of investing is that transaction fees tend to eat up your profits. My brother, who works on finance, placed a bet on gold prices with some of his personal savings; his prediction was correct, but he lost money on the transaction fees.

To avoid the mentioned risk-seeking behaviour, you could split your $50 proportionally to the amount of virtual money made.

Doesn't this incentivize people to convince other people to adopt bad investment strategies? To actually incentivize everyone trying strategies that maximize expected value simplicio could award each player some fixed multiple of the money they make (maybe up to some reasonable cap).

It does. I haven't been considering communication among players.

The problem is that you're setting up the goal of being #1 in a (presumably) large group, which creates wildly different incentives than actual investing. The proper strategy is to pick a single investment with volatility such that it has a (1/number of participants) chance of being utterly dominant at the end of the period, because a moderate return is no better than bankruptcy.

This is a good point. I think that part of the plan was a bad idea.

I've done well at stock market speculation over the past 12 years. It took me 20 years of doing it to become good at it. If I'd had LW available when I started I expect I could have become good in around 5 years.

One of the difficulties was that with most strategies, 1-year returns are sufficiently dominated by luck that the feedback is nearly useless.

The contest might be valuable if it lasted close to 5 years and wasn't a winner-take-all event. Would that attract any participants?

I'd be interested.

To what skill sets do you attribute your current success?

One of the troubles with stock contests is that the timescale becomes very relevant. I like that you've picked a year- most of the ones I've seen are only a few weeks, and so dramatically overreward luck and leverage. Even then, a year is not all that much time for the value investor (the brand that tends to make the most out of not sharing Mr. Market's crazes), and I suspect you're better off learning investing from investing professionals (either by getting an account with Motley Fool, or reading The Intelligent Investor, or so on).

I remember reading a few years ago about a fund that wanted to use data mining of investment newspapers for psychologically relevant words, to predict how Mr. Market would behave, and so make money off of anticipating his swings. I don't remember its name and don't seem to be able to find it now.

Have people put in real money. That way incentives are aligned. Anyone who can't afford to put in a few grant has more important things to be doing than studying the stock market anyway (exception: investment banking interns)

The source of that story is Benjamin Graham.

http://en.wikipedia.org/wiki/Benjamin_Graham

In general I suspect that a) EMH is wrong in some cases b) Good amount of instrumental rationality should give one some advantage (Warren Buffet and his partner Charlie Munger have studied rationality, including specifically Kahneman).

I'm not sure having a contest is the best way to go about this though. I've tried to convince some people (unsuccessfully) to engage in a group study of the area, my goal was to create a "portfolio team" where expertise may be shared and people can better control their biases.

Sounds interesting! My participation is primarily contingent on how easy the software platform you end up finding is to use.

I once participated in a contest very similar to this one against a class of upper level engineering students. It turned out that I couldn't use the software package required for the contest due to irrelevant VPN issues. After the contest period was up, I had my $5 of play money sitting untouched, and every other student had less than $5 due. I won the contest.

Just sayin', being able to use the software is not necessarily required to win.

Assume this guy is right. How long before the "low volatility" premium disappears?

If I actually had a good way of picking stocks, I would simply invest in those stocks. The $50 is hardly worth bothering with.

If I wanted to test whether I had a good way of picking stocks, I would be hesitant to do it with real money. I had admittedly very vague plans to test various ways of picking stocks with play money on my own, but those vague plans are more likely to coalesce into action if simplicio's proposal actually happens.

Several free stock market simulators already exist.

Hmm. I really should have tried searching for that. Thanks!