building a pretty good prediction markets trading firm
Joined May 2026
- Tweets390
- Following122
- Followers735
- Likes766
the white pill on "very fast but not fastest" hft infra is it's mostly commoditized by a few vendors who compete on price. many smol teams, or even a few exceptional individuals, can afford it.
but if you want to do low alpha, ultra low latency hard arb, it's a different story.
i kid you not lads, in 2021 you could blindly quote alts 200 bps wide with 300 lines of python and make 7 figs
there's infinite alpha in being ok at socializing as a quant trader, way more than being marginally more cracked.
the bar for social skills is so low that the marginal returns are huge.
2.2M/4 yrs+385k for an engineering lead. So $935k. Junior eng role is worse. Idk why quants are all thirsting for these lab jobs, you can get this in all cash at a tier 2 HFT shop.
Maybe the superstar researchers have mega millions comp, but not rank and file devs.
ok since everyone is telling me about equity appreciation
if you joined now you'd be looking at 1.3 trillion or whatever they price it at, if appreciation was guaranteed, investing would be the easiest business in the world
also when oai ipo's (soon) you can just take your cash comp and buy the stock if you want, or buy a different stock you like better!
alright
tier 1 jane st, hrt, citsec, xtx, jump
tier 1.5 optiver, imc, sig
tier 2+ literally everyone else
inb4 "but akshually rentec and tgs" idc they are boomer scientist places who don't really hire kids for QR, kinda true of xtx also tbh
sorry if im missing anyone
really confidence inspiring that whenever you deal with a crypto company the wire info is like Mule Barn, North Dakota (population: 3) Community Bank
alright which farming operation is not even bothering to give their subaccounts plausibly unrelated names when there's a public leaderboard? cmon guys, this is amateur hour.
are you gonna mention Polymarket washtrading through same name accounts?
i found 24 "plys" accounts created around the same time that appear to be farming volume on perp DEX
their volume makes up 18% of ALL-TIME Polymarket Perps volume
you can open the leaderboard and verify the accounts yourself
good enough for the Token2049 announcement?
The bloomberg terminal has such a complete cultural victory as a status symbol.
You can be worth a billion and still feel compelled to flex your $2700/mo financial software subscription.
On prediction markets and sports -
When I set out to trade prediction markets, I was really excited about the truth machine thesis. I wanted to trade politics, global events, etc, and contribute to the probabilities on the screen.
But running a business is a survival game, and to survive you need to meet the demands of your counterparties. And the demand is in sports.
Am I thrilled to tell people I'm a pro sports trader? Not really tbh. Is the pnl good? Yes.
I suspect some of the early employees at kalshi and poly feel the same way, based on what their businesses set out to do, and what they have become.
True. I grinded jane st interview prep like the single most important thing in life, hundreds of hours, compiling every post on chinese forums, trying to program simulators of the interview... Pathetic I know. Rejected final round.
Fuck 'em, printing money in this game anyway.
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
sure the user is getting a bad price on the ultra longshots, but they kinda suck for the mm too
assuming all 0.1c are worth 0 (they're not), it's optimistically 36.5% ROIC if all expire in 1 day (they don't).
adjusted for toxicity, might as well just put your money in the s&p
Extremely long parlays, which have become an ever-growing part of Kalshi’s activity, are creating by far the biggest losses for Kalshi takers, with the prices of the longest combinations more or less completely detached from their probability.
That the biggest market maker edge happens on the longest parlays may not be a surprise, but the extent of the difference is stark.
