Lebed, Redux

Plus: "Entrepreneurs," egregores, software as modern art

Programming note: I’m going to start doing more posts on Substack rather than Medium, and will be adding a subscription option starting March. Friday newsletter updates will remain free; I’ll do two or more additional pieces per week for subscribers.

Lebed, Redux

One of the best winning streaks in financial history belongs to an independent prop trader named Jonathan Lebed, who made roughly 10,000% from late 1996 to early 2000. Lebed was an agile trader with great timing, especially in small- and micro-cap stocks. He was also, in 2000, fifteen years old.

Lebed’s strategy is very well-documented, because the SEC made him stop: he’d buy stocks, tout them on Yahoo Finance message boards, and sell. What you think of this approach depends on what you’ve read about it. The Michael Lewis article is a good start, arguing that Lebed’s behavior was, in substance, identical to what analysts and amateurs do every day: everyone touts stocks they want to go up, and by induction many of the most energetic touts have serious skin in the game.

But read more, and a darker picture emerges. Lebed didn’t just buy and promote; he bought thinly-traded stocks in size close to the end of the trading day. The basic model was:

  1. Lebed buys in size at 3:58pm.
  2. Stock rises 10-30% on 10x average volume.
  3. Investors look at message boards to figure out what’s going on.
  4. They find Lebed’s writeups, full of ALL CAPS exhortations, hints of inside information, and a high price target.
  5. For some reason, they buy. One of the sellers is one Jon Lebed.

You can dig deeper. The forums are gone now, or at least very hard to find, but back in the early 2000s I got interested in this case and read the old Yahoo Finance and Raging Bull forums where Lebed plied his trade. Michael Lewis notes that the SEC tried to find a link between Lebed and a stock promoter named Ira Monas. Monas was an energetic crook: in 2001, he was busted for running a financial scam from prison.

There’s ample circumstantial evidence on the message boards that Lebed and Monas were in close contact, and that they worked together on several of these market manipulation schemes. Monas was an experienced pump-and-dump operator, and it’s likely that Lebed either learned the trade from him or literally took instructions from him.

Either way, it’s not so much a story of a precocious kid who enacted the reductio ad absurdum of CNBC and sell-side research, and more of a sad story combining everyday market manipulation with a very depressing violation of child labor laws.


Wall Street Bets is a reddit community devoted to goofing off about finance. Goofing off as a high viral coefficient compared to discussing stat-arb models or parsing earnings call transcripts, so it naturally has a huge following—almost 900,000 readers. And thanks to decades of Silicon Valley UI/UX research, it’s easier than ever to speculate on stocks and options.

You know where this is going.

Yesterday, someone posted a bullish writeup of Lumber Liquidators to WSB. The stock rose 18.6%.

This feels very postmodern. The whole community is in on a joke, but to really participate you have to actually day-trade on crazy ideas. Putting 80% of your savings into call options is the WSB equivalent of having ashes on your forehead on Ash Wednesday.

The LL writeup is not terrible, at least for a retail investor—a professional would not talk about purchases from indexers like Blackrock and Vanguard as a bullish signal—although the suggestion to buy short-dated calls is bold to say the least.

What’s really interesting here is the high likelihood that more organized groups will learn to exploit this. It happens in other fields; street fashion turns into regular fashion, at a high markup; indie music turns mainstream; Mike Bloomberg apparently pays staffers extremely well to make Mike-friendly memes. But in equities, the scalable professionals are either a) people who work in the industry and are constrained in how much they can tout, and b) criminals. People reading WSB think they’re reading a Lebed—a goofball who exaggerates what normal professionals do. But they might be reading a Monas instead, someone who has made a career out of bilking investors.

I don’t have any way of knowing which of those categories the Lumber Liquidators poster falls into. But if I can see an opportunity for market manipulation, I’m sure a professional stock promoter can see it, too. Be careful out there.


Digital Ocean has raised $100m in debt financing. This fits in with the thesis I mentioned in yesterday’s post: companies with knowable unit economics should borrow, not sell equity.

Every so often, the Kaufman Foundation will release a study saying that startup founders are, contrary to popular opinion, mostly middle-aged, not young. And every time they do, I’ll internally roll my eyes and say “You’re counting real estate agents, accountants, and lawyers as ‘startups,’ aren’t you?” The answer is yes. Brian Timar has approached the question with more rigor, taking a list of influential startups and checking the founders’ ages; his histogram is close to the conventional wisdom. Statistical claims like this aren’t action-guiding in any meaningful sense; the 60-year-olds who quit their jobs to start new companies are a very different demographic from the ones who don’t (Colonel Sanders, Cornelius Vanderbilt, and Sherman Fairchild all got into the businesses they were famous for at advanced ages, for example). It’s mostly an argument that if someone makes a surprising statistical claim, it’s probably because they used an unconventional definition.

Quinn Norton has an extended meditation on what it’s like to get Canceled. Internet culture has reinvented certain aspects of Christianity (original sin) and forgotten about others (confession, forgiveness, redemption). There are plenty of examples in biology of a simplified version of an organism thriving in a new environment, but most of them involve extremely unpleasant viral infections and intestinal parasites.

Michael Gibson points out the paradox of big tech and politics: major tech companies seem powerful at a national level, but they can’t accelerate the multi-decade process of building new housing units in the Bay. This might be to their benefit; my Peak California argument is that mature tech companies can afford higher cash comp, pricing out smaller companies that pay equity. But it’s probably not a cynical ploy: tech companies are all about big abstractions; when they say “local” it usually ends in “-ization.” Fixing SV’s zoning issues is just not something Google can scale to a billion users, so they won’t even try.

A while ago, I wrote about partisan media as a de facto electoral college. A friend pointed to this essay, which argues that they’re not electoral voters, just gods.

This interview with Sequoia’s Doug Leone is great. Highlights include Sequoia’s painful quest to get a positive IRR on a fund that invested at the peak of the dot-com bubble, and a great line about the importance of early investing: “If you lose seed in venture, you become (as I say) a private equity firm because later on, all you have to compete is on price.” This lines up with my argument about VC skill: that the reason to invest early is to get the opportunity to invest on favorable terms later on. In a world where companies can stay private up to valuations of $50bn or more, seed and Series A investing without follow-ons are, increasingly, a hobby.

And finally, Alex Danco connects the techlash and postmodernism. Read the whole thing. He highlights a Google Ngram search shows that mentions of “progress” were surpassed by mentions of “innovation” in the early 1970s, another one for the “WTF Happened in 1971?” files. The real meat of the essay is the question of whether software is building the future or optimizing the present. Software is just a way to take a stream of symbols as an input and produce another stream of symbols as an output, according to particular rules. It’s valuable when these symbols represent truths about the world, but there is something deeply depressing here. A friend describes modern art as art for people who are better at arguing than drawing; if progress only happens in software, it’s all argument, no action.