Sabitlenmiş Tweet
@thodoha @therobotjames @laurentzeimes @macrocephalopod MaxDama booklet and @AgustinLebron3 book contain 99% of what is really needed.
English
Martin Gale (λ,π)
409 posts
Martin Gale (λ,π)
@itsmartingale
Former Head of Monte Carlo simulations at @ARKinvest
SO(n) Katılım Mayıs 2020
144 Takip Edilen1.7K Takipçiler
Chart crime 101
Noel Smith@NoelSmith
It's cool to pay attention to bond volatility (makes Fonz sound)
English
@macrocephalopod @elletwocache Interpreting it as “Alpha Impact” makes the formula correct (or very close to)
English
@sdav1986 Also very much not clear what is “revenue” in a pass through fee pod shop.
English
@famadeo Tradfi. Last time I traded crypto was 2017 in my PA
English
I'm off work tomorrow. Ask any questions below and I'll answer while at the airport waiting for my flight.
English
@macrocephalopod Dynkin lemma proof is a long time favorite
English
@itsmartingale What’s your recommendation for the best book for a second course on probability?
English
@RootTwo23683 @VivekVRao1 Having done this in real life for a long time helps in filtering out garbage papers from good ideas. Main red flag with this paper is that neural network works amazing and simple signal does not. In reality simple stuff works ok.
English
@itsmartingale @VivekVRao1 How can you tell it has some selection and look ahead bias?
English
I think the paper cited is Statistical Arbitrage in Rank Space by Li and Papanicolaou papers.ssrn.com/sol3/papers.cf…. "In this paper, we show the superiority of statistical arbitrage in rank space over name space by exploiting the enhanced mean-reversion of residual returns in rank space. Our statistical arbitrage portfolios obtained using neural networks in rank space achieve an average annual return 35.68% and an average Sharpe ratio of 3.28 from 2007 to 2022 with 2 basis points transaction cost. This contrasts with the conventional statistical arbitrage in name space that yields negligible returns during the same period."
sysls@systematicls
If you join a statarb team, you are essentially trying to run variants of mean-reversion (MR) strategies. The catch is this that traditional MR (betting that gross returns will revert) is negative after-cost since 2010. So, how have statarb teams adapted? Simple, they've learnt how to make mean-reversion strategies in a different "space". "Gross returns" are an example of a statarb "space". Today, i want to talk about a really novel "space" that still produces meaningful performance even till today. Perhaps some of you may have a clue, but the hint is that it involves rethinking factor decomposition! If you'd like to find out more, head over to the bad place AND as usual, I will give out random free reads to retweets + comments on THIS POST.
English
@quantymacro @VivekVRao1 I've not read any interesting paper recently tbh
English
**Request to all portfolio managers**
I am negotiating the 2nd ed. of “Advanced Portfolio Management” with Wiley. It will be better: error-free, new material, AND IN COLOR! (If not, I will rewrite and publish elsewhere). Here is the request: what topics do you think were missing? Answer in replies. If your advice is accepted you will be acknowledged in the book.
English
>leverage AI to test an obscene amounts of trading / pricing parameters
>best results are still a couple EWMA's in a trenchcoat and a linear skew on position
English
City of Milan really managed to dig a new underground line for the city airport of Linate and not have it stop at any of the five main train stations. Impressive
English
HFT developer interview question.
What do you think this represents?
Is there anything wrong with this latency distribution?
English
@itsmartingale Timer threads evicting cache, interrupts not steered properly, etc.
English
@KlondikeFX not a bad guess: making sure the measurement tool is accurate is something lots of people forget about
English
@itsmartingale Possibly a long tail, and the spiky shape may reflect clock resolution or measurement artifacts.
English
@itsmartingale Allocation far right, syscall middle, pipeline stall in left, this is probably getting picked off because the trade path has too much jitter
English
@maxsparr0w spoiler: this is not exchange latency to t2t timing
English
@itsmartingale First thing to clarify is how this was measured, is this from exchange to wire, exchange timestamps might have their specifics including clock sync, also is the measurement point fixed and reading wire time or it's aggregated across many points
English