I’m no expert in the space but was under the impression FTX was still considered a low volume exchange. If Alameda mm was still providing substantial volume they may have felt the price they were paying being bad at trading was worth it from a marketing point of view. This was the same crew buying stadium naming rights and over paying for eSports teams.
As for hiring an adult, have you seen some of the things they’ve said in public? They thought the adults were wrong! Like on a mathematical level. People that far into their own cult-think don’t want outside help.
Finally, the Ontario Teachers thing feels extremely overblown. They put in 75m of ~220b aum. That’s a minuscule investment in a category that was clearly part of their highly speculative book. I think people keep bringing it up for the “think of the teachers!” benefit, but of a pension fund doesn’t have some outsized return opportunities on their book they’ll not be able to cover their outflows under certain future economic scenarios.
OTP put in something like that in the fund I was at. Yes it's small for them but it's still serious, the guy in charge wants to move on to bigger things.
Oh and I totally forgot to comment on the math. Kelly criterion, I don't get how the guy could have studied physics at MIT and not understood it. It's simply a result that tells you how much to bet of your stash if you are presented with some repeated betting opportunity. It's essentially an extension of high school level probability theory, with similar clean-room constraints (independence, etc). There's also a continuous version of it using Sharpe ratios, no surprises.
If you read anything about it, it's very very not smart to bet beyond the ideal size. This is actually the reason why a lot of people use less than the ideal, because you don't want to end up over the hump. Both the assumptions that underlie the result and the measurements of the parameters are not easy to ascertain, so don't act like it's a dice game within a known environment.
I don’t want to besmirch this particular person I’ve never met, but I definitely have met engineers who look at the stock market, look at their sophomore or junior level signal processing class and decide that all those Econ guys are a bunch of dummies and they could totally beat the market with some Fourier transforms or whatever (the fact that investing firms hired a bunch of quants is left unaddressed). So, I could see somebody making this sort of error.
I'm friends with a retired quant, and one thing I learned from him is the big firms on wall street will hire the absolute most brilliant people they can even to work on utterly banal well known basic strategies. PhD's doing high school math. Because when you're moving billions, nothing is actually basic.
Kelly criterion is not the best strategy for one. It’s the best strategy for an ensemble of ones in many universes but you can only live in one universe. It’s a nerd trap because it seems so elegant and reasonable.
This isn't true? If you aggregate utility across universes using addition and have utility linear in wealth then you maximize your aggregate utility by going all-in on every favorable bet.
Kelly turns out to be equivalent to maximizing expected log-wealth, if one stays within a single universe, or to maximizing the probability that one will eventually have more wealth than any agent in the same universe that employs any other strategy on the same sequence of opportunities.
This is the worst comment I have ever read. There are reasonable definitions of expected value and probability that do not require infinitely many universes to figure out whether you should raise pre-flop with aces.
You are conflating time-samples with ensembles. That lecture refers to finite time series which may be shown to converge for N < infinity. That’s fine.
I have not mentioned either of these things. You can find other finitist formulations of probability if you want.
You replied to a post about aces. For any finite number of chips and players (n_players > 23 starts to raise some questions), there are finitely many games of NLHE anyway.
According to a random source, they moved volume on the order of 500 billion. Transaction fees of 0.1% for 500 billion is a lot of million to keep the lights on.
For a maker and taker with best fees, the exchange only made 0.5bps, not 10bps. We happen to know that top rebates were pretty hard to achieve though, so most volume probably paid at least a couple more tenths of a basis point.
In the trailing 24 hours, presumably sampled at the time FTX's servers for dynamic content such as logging in and trading were shut down. also did you reply to the wrong person?
Yes, the crash at FTX gets all of the press, but it's really the slow motion collapse of things liken pension funds that should really be getting our attention. SBF is just a piker compared to the pension funds.
As for hiring an adult, have you seen some of the things they’ve said in public? They thought the adults were wrong! Like on a mathematical level. People that far into their own cult-think don’t want outside help.
Finally, the Ontario Teachers thing feels extremely overblown. They put in 75m of ~220b aum. That’s a minuscule investment in a category that was clearly part of their highly speculative book. I think people keep bringing it up for the “think of the teachers!” benefit, but of a pension fund doesn’t have some outsized return opportunities on their book they’ll not be able to cover their outflows under certain future economic scenarios.