semaji.eth

Formerly HFT QT. Used to write at Quant trading, prediction markets, crypto.

[semaji.eth @semajeth]

https://t.co/CStnZNax2N

[semaji.eth @semajeth]

@notsethbregman It is highly immoral of you to provoke him to bury his foot into his whiny little runt of a dog. You are effectively instigating animal abuse.

[semaji.eth @semajeth]

@yenwod_ This was one of the best resources posted this year, hope the random uninformed sneers about it not being true HFT (made largely by people who think crypto runs on ASICs and nanos) did not land.

[semaji.eth @semajeth]

@TheFatMbappe @IvanCryptoSlav @shayne_coplan Would be interested to know roughly how you know, but appreciate it is likely quite confidential

[semaji.eth @semajeth]

@christinaqi I appreciate that it may be difficult to do this publicly, but I would be supremely interested to see the before-and-post-revamp decks and what the difference concretely looks like.

[semaji.eth @semajeth]

@NTmoney ...I could go on, but you get the point. Ten good people could turn things around for Poly, but the idea that the same personnel who have ran a generational company into the ground will now "lock "in to turn it around, with no new blood driving higher standards, is total cope.

[semaji.eth @semajeth]

@NTmoney ...which, absent of change, will continue. Poly has bad comms with its native userbase (many of them on X), its resolution issues are ubiquitous, its technical design remains meh, has a culture of unprofessionalism (I have personally experienced this), has ZSC-type scandals...

[semaji.eth @semajeth]

@SamGreenwoodRIO In general that is a more reasonable prior than that of "my opponent is trying to exploit me back" which requires your opponents to a) have enough data to definitively know where your imbalances are and b) understand specifically how to counter them

[semaji.eth @semajeth]

@GTOWizard Though it is ironic that this is the account making this post, as GTOWizard is almost synonymous with combo-level solver analysis!

[semaji.eth @semajeth]

@GTOWizard ...and are broadly unaware of why they're using b33 or how their turn strategy should deviate from solver when V underraises flop by 50%, of how their strat + adjustments ramify down the game tree - while being comically aware of microfactors 100x less impactful on overall EV.

[Sarah @sarahskye]

@nayibbukele And as you said:

[semaji.eth @semajeth]

Various people getting railed in the comments. The most common misconception here is that it depends on how many cards have been turned over - but this is a boring (and trivial) question and not what's being asked. We're averaging across all permutations to find an a priori!

[Tomek Czajka @TCzajka]

A standard 52-card deck is shuffled. You turn over cards from the top, one at a time, until the first ace appears. What is the probability that the next card is also an ace?

[semaji.eth @semajeth]

@mbeisen Not really sure why people are going at you in the comments. I agree that this is a very cool problem because instant intuition is correct, but there are very complicated (and wrong) reasons to doubt that intuition and clearing out those reasons is difficult and fascinating!

[semaji.eth @semajeth]

@avi_eisen (The conclusion doesn't follow! But I have to be honest that in an interview I would have gone with this)

[semaji.eth @semajeth]

@avi_eisen Ignore all the cards in the deck except 2c and the 4 aces. The probability of the 2c being picked first is 20%, (just uniform) and its complement is 80%. Now consider only the 4 aces: the probability of the first ace being As is 25%, so its complement is 75%. So 2c is likelier.

[semaji.eth @semajeth]

@euler_mab Would love to hear!

[semaji.eth @semajeth]

@CGkol This looks very cool! Two questions came to mind for me: a) what differentiates Metric from Uniswap? If it's price discovery, what makes the decoupling better than Uni's dual mandate? b) very concretely, how does this avoid the typical AMM adverse selection issue?

[semaji.eth @semajeth]

@waterloo_intern @jurlycat GTO alignment coherent extrapolated volition, if we're being precise