mbappe not winning the balondor is so controversial
that's what an attention seeker would say, now i have your attention
let's talk about how a market can put a price on arguments like this, in real time
it's called LMSRs
most prediction markets you use run on an order book
but there's another way to run a market, and it's called LMSR which means logarithmic market scoring rule
it was proposed by economist @robinhanson , and the idea is simple
be ready for a lot of maths :)
what if a market maker was always ready to trade, with a price that moves as people buy?
with no waiting for someone to take the other side
here is how it works:
C = b × ln( e^(q_yes / b) + e^(q_no / b) )
⤷C is the cost function, that's the market running total
⤷q is how many shares have been sold on each side
⤷b is a setting that decides how deep the prediction market is
you don't pay a fixed price, you pay the difference in C before and after your trade
let's use b = 100 and a market that starts at 50c
⤷ you buy 50 YES shares and pay $28, the price of YES then moves to 62%
⤷ you buy 50 more and pay $34, because the price is now higher, YES now moves to 73%
your own buying pushes the price up and that's slippage
now the part I like most, the market maker can never lose more than b × 0.693
why? the cap exists because of the maths
the formula is built so the market maker's worst case loss is b × ln 2, no matter how much anyone buys
with b = 100, the most it can lose is $69
the person who creates the market puts that in up front, and that's the whole risk
so b is a trade off
⤷ buy 200 YES at 50c with b = 100 and you pay $143, and the price jumps to 88%
⤷ buy the same 200 with b = 1000 and you pay $105, and the price only moves to 55%
a big b means a deep market, low slippage, but the creator risks more money
a small b means a thin market, big price swings, but little risk for the creator
so the next time you see a price jump on a small market, ask yourself how deep it is
in LMSRs, the depth is just a letter "b"
thanks for reading :)