The Research Desk: A First Bet on Betting
Betting markets and the case for taking a punt on them.
Welcome back to the alternative desk. It’s time to introduce one of the types of posts you’ll be receiving from me. The Research Desk.
The intention is for these posts to cover everything about the novel strategies we’re attempting to find; their universes, their hypotheses, their execution and the inevitable practical issues that come with trying to find edges in more exotic environments.
With that being said, let’s introduce our universe for this quarter:
BETTING MARKETS
The Beginners’ Guide to Betting Markets
Underwhelmed? Thanks. Don’t be deceived simply because we’ve not decided to rival Jane Street in Q1. 10% of the Premier League is majority-owned by individuals who’ve made their fortune from sports betting. There are people making some very serious money in this space.
Before we delve into the wider world of sports betting, it’s important to note that the fundamentals of profitable betting rests on one concept: expected value (EV). To understand what’s below, you only need the last sentence of the next paragraph. To understand what’s coming over the quarter, we’ll have a specific first principles post dedicated to the topic of EV coming in the next newsletter.
Every line on a market implies a probability. The tidiest mathematical example is a bet priced evens (2.0, +100, 1/1, depending on your odds preference). A price at evens means your bet needs to win 50% of the time to break even. If you think it’s likelier, you should bet; otherwise you shouldn’t. Maybe you think that in reality there’s around a 45% of the bet winning. Then you don’t bet. But it isn’t only your view that makes a bet profitable or unprofitable, if you can find the same event priced at improved odds of 2.5, breakeven drops to 40% (the reciprocal), and the same bet is now worth taking. Fundamentally: lower odds are always worse for the bettor, so where you bet, and how competitive their odds are, matters.
Betting companies exist all over the world, from Asia, to Europe (where this newsletter is based), to now much more prominently in the US. The premise for the completely uninitiated is relatively simple. Sign up, deposit funds, find a bet that takes your fancy, and lose. At least on average. Somebody needs to pay for the shirt sponsorships. And yet where you bet can matter as much as what you bet. Almost all platforms fall into one of three categories:
Soft Books
See: Sportsbooks, Retail Books, Bookmakers
Examples: Bet365, Fanduel, BetMGM, literally thousands of others
Associated Terms: punters, casual betting, losing long-term, my parlay only lost by one leg
What makes a soft book? The business model for these is simple; offer bad odds so your customers lose quite a lot of money on average (refer back to EV paragraph if this relationship is unclear). Spend heavily on marketing, sign-up offers, sponsorships and brand visibility to attract as many casual bettors as possible to your brand, as opposed to any other sportsbook. These are the most visible and well known betting platforms, specifically because their business relies on it.
If we hypothetically could find a winning strategy, would this work well for us? No. These platforms have a rather pesky habit of outright banning any customers who manage to win long-term. There can be a delay on the banhammer, meaning there’s potentially some fun to be had in the interim on soft books (search matched betting), but it’s more of a summer project than the type of long-run quantitative strategies that we’re aiming for here. Unfortunately I’ve already completed this summer project, in several names, over several summers, and remain banned from many of these soft books. So these are a non-starter.
Sharp books
See: Asian books, Low-vig books
Examples: Pinnacle, Circa Sports, SBO Bet (still many but probably <40 true sharp books)
Associated terms: Syndicate betting, high-limit betting, low-vig, brokers
Sharp books, Asian books, or any of the derived terms, run close to the opposite model. They offer highly competitive odds, deliberately take high-limit customers and syndicates, and welcome winning players rather than banning them. They also manage their prices aggressively, moving lines in response to the action they take, including when sharp accounts pile in on value. It’s the closest thing betting has to price discovery in a financial markets sense, and the closing line on a sharp book is widely treated as the truest available estimate of an event’s real odds.
Would a winning strategy work here? Yes and no. The catch is jurisdiction. Sharp books tend to be based in Asia, and actually getting money on their system and on their lines takes a fair amount of monetary plumbing for a European. They often also come with expectations around stake volume that a four-figure, experimental bankroll can’t meet, which makes them unfeasible, or at best sub-optimal, for us.
Betting Exchanges
See: Peer-to-peer books
Examples: Betfair Exchange, Smarkets, generally more specialised
Associated Terms: Liquidity, Back/Lay, Am I the dumb money?
Betting exchanges operate somewhere between the sharp and soft books. Or maybe not between them, but on a slightly different axis altogether. The fundamental difference is that everything here is peer-to-peer. When I back (buy) a bet, there’s someone on the other side laying (selling) it, effectively short that same outcome. Their payoff is the mirror of mine: if the bet loses they keep my stake, and if it wins they pay out my winnings. The market parallels run deeper than that. Each line has a spread, the best back price you can get sits just below the best lay price next to it. Taking a bet right now means crossing that spread, exactly like paying the bid-ask to trade a security immediately rather than waiting for your price. If you want to be filled without hanging around, you pay for the privilege.
As for the business case, exchanges take a set commission on net winnings, normally 2-5%. They don’t set prices themselves, but a version of the same price discovery happens here, often more transparently, because most exchanges show the order book. And of course, the incentive to cap winners evaporates; they want your volume so they can get the commission. The catch is liquidity: your bet only exists if a peer is willing to match it. The biggest players still prefer sharp books for the high limits, so the sharpest action happens there first, and exchange market makers tend to move their prices in response to it. As for who provides that liquidity, it tends to be large syndicates and professional market-making groups. Outside Betfair, these groups can be connected in some way to the owners of the exchange itself, which I’m told regulators love.
But how well do exchanges suit our needs? For the most part, very. I have access to several exchanges in my jurisdiction, and better still, most run APIs, so odds can be pulled, processed and bet on programmatically. Of the three venue types, exchanges fit the brief by far the best.
The plan from here
So where does this leave us? If we’re going to find the consistent, scalable and algorithmic edges we want then betting exchanges are the place to look. The job is to find inefficiencies that show up in these markets, confirm they’re real rather than noise, and turn each one into a strategy that runs itself. Once a strategy is automated and behaving, move on and start hunting the next, while the last one keeps ticking over in the background. Repeat until we have a stack of small edges running in parallel and buy a Premier League football club. Simple.
After we’ve taken our deep-dive into EV, we’ll start unglamorous and venture out arb-hunting, to prove that our plumbing works before delving into the really niche stuff. See you next time.
The Alternative Desk
Still a going concern