hedge fund quantitative macro tradooor. chief ridge officer. writes some (allegedly) advanced cool quant stuffs on my blog. research articles in highlights tab
the start of my grifting era.
yes I'm starting a "newsletter".
but who the f am I?
I completed my undergraduate last year, & received multiple offers from PMs in Citadel/MLP/Bluecrest/Rokos/Brevan to join their pod w/o going through the typical grad rotations. and I did that
this vid appeared on my YT page and brings back memories. it came out during 2020 lockdown, where I decided to go all-in to pursue quant finance. didn’t know a single thing about coding or regression back then. now to think that I’m doing it for a living is insane. super grateful
It is with great sadness that the Simons Foundation announces the death of its co-founder and chair emeritus, James Harris Simons. Jim was an award-winning mathematician, a legendary investor and a generous philanthropist.
me reading this: “damn even XTX is only paying PhD fellas 40k/yr poor all the PhDs always getting underpaid”
“ohhh….. it’s 40k/month…. base salary….. 3 days a week….. how long is a PhD again where do I apply….”
what tf is “Renaissance Technologies”? never met anyone who worked there. they use simple regression lol ngmi for sure. even in the case of univariate regression, Ridge Regression is *guaranteed* to be better than OLS as long as ridge penalty < 2σ^2/abs(β). founder’s a LARP.
he was just like me fr (except he was a billionaire and my net worth is negative on a good day). thank god my PM shares the same disdain towards brainteasers/olympiad qs/mental maths bcs wallahi I can't multiply 2 digit numbers to save my life
now I’m actually living the *exact* life that I dreamed about few years ago. I work in my dream firm, in my dream role, and live in my dream city. back then I would’ve given up my thumbs if that’s what it would take.
my finances are great. before this I would eat a heap of
girl at a Penn station food stall calls me “my love”, then starts asking about my guitar, then calls me sweetie when I go to pick up my sandwich
what could she mean by this
the secret to successful pairs trading is not cointegration or stationarity its actually quoting your correlation values to at least 16 significant figures. Citadel has a special C++ team to quote it up to 128 significant figures. few know this.
If you don’t understand calculus and differential equations , you should give up trading
This is what it takes to generate proprietary trading signals to live luxurious life
There is no substitute for hard work
I know two discretionary macro PMs both ~40y old learning Python on the side. and some of you busy bookmarking 102828 books but never even finished a single tutorial. gotta pick it up my brothers. if not next year you’ll be at the same spot wishing you did start today.
if you’re 23 and saying it’s over because you didn’t start at 16 or something you’re a loser looking for an excuse to not do anything and you’ll get beaten by people learning ML in their 30s
Long time lurker, finally decided to engage with the community and put my content out. Here's a post on FX mean-reversion strat, and thought process behind it. Very much influenced by
@macrocephalopod
@investingidiocy
(quoted them many times in the post)
“All these hedge fund quants are not actually researching strategies. They sit at their desks and throw bunch of shit just to end up doing momentum and mean reversion. And if they get negative results they just multiply by -1. Its definitely true bro you can look it up its crazy”
2020 I did my first Python project. not quant finance but on bulldozer. my maths was bad. I didn’t know the diff btw covariance/correlation
I wasn’t thinking abt Citadel or MLP. I emailed local banks asking to work for *free*.
everyone started somewhere. dream big and go get it
this is not surprising at all. given the number of HF managers, Senior Quant Top G Researchers, Head of HFT Big Dawgs casually dropping 927372 textbooks/papers/github to read here, we have so much alpha just laying around. it’s up to you to make something out of it. Choose Rich™️
trying to find good books? sick and tired of pretentious quant books that are too mathsy and not readable?
dont worry i got you. who said rates trading concepts cannot be written through a story of an immigrant? Fabozzi is really one of the most creative minds in finance.
while I’m on the topic, if you only can read one “how to land a quant role” post, it should be this one. this should answer 99% of DMs I get. and pls read the last paragraph slowly and carefully. and read it again. and again.
look at all the pathetic souls on a Sunday afternoon. mindlessly strolling around Regent’s Park like sheep taking pictures of flowers. they don’t have even the slightest idea that Ridge Regression dominates OLS for a range of values even when there is zero multicollinearity.
every single day I can't help but to wonder how did I finesse this firm to trust me to write code. turns out you can do useful things by just being mid at pandas & numpy. don't let the jupyter notebook haters intimidate you. we the jupyter notebookers are THRIVING
#silentmajority
Friday quant-fund janitor intern interview questions:
1) find the eigenvalues of the covariance matrix
2) describe the behaviour of the ratio of the larger to the smaller eigenvalue as rho varies
3) what insights do you get from this?
I did some work for a Quant Researcher-turned-macro PM, and my code was described as “average”. I also did work for a discretionary macro PM, and was told that I’m an “extremely good coder”. so if you’re mid, pick your battles. in the land of the blind, the one-eyed man is king
my gym PT told me he’s been learning Python for 2hrs everyday. today he was excited after learning f-strings. he didn’t even go to uni, just wants to learn because his dream is to create a fitness app. and most of you still haven’t watched 2 hour tuto video you bookmarked 3m ago
you might think it’s fun to be a junior quant trading butterflies. but flies are notoriously hard to trade. at inception it’s long Volga, flat Vanna, long Sega. but it’ll change as spot moves. and even if you somehow make money, you’ll go home & find yourself short Vulva. tough.
tally count of terms I’ve heard in the office since joining a HF:
fat tails : 0
non-normal : 0
cointegration: -9999
beta/regression : 642
why is the curve marked wrong: 437
tranquilo brother let me fix it: 437
its a 1st gen exo dont book it as a 2nd gen: 1
it wont work: 8711
a cool article about backtest:
have you ever wondered why practitioners run regression as a pseudo-backtest? (one) reason is backtest ≈ predictive regression. this is only 1 page, the full book is a must read IMO.
Inefficiently Efficient - Lasse Heje Pedersen
I like reading Two Sigma’s job postings. the job responsibilities feel more descriptive and specific vs other typical ones. also I’m not super familiar with all of the things on the list, so I also get to learn something + some ideas from it. quite cool tbh
Day 15/30 of studying regressions *only*
reading hits differently under the sun. might put flaneur in bio and start writing a book about fat tailed distributions soon. follow for more updates.
probs more True than it is False. I grew up in a country where there was 0 (yes, 0) hedge fund. to say I felt extremely behind when I first found out what a HF was would be an understatement. from my seat I can see ~35ppl, and none of them (yes, 0) has similar background to mine
not joking if I win $100mm I would pay the smartest of the top dawgs to teach me maths at my home. I would go through maths papers/books & literally ask them to explain it line-by-line. extra money if you teach me how to derive optimal ridge penalty for anisotropic features.
stupid story:
I used to have Citadel as my phone’s wallpaper (before you mfkers judge me for being a tryhard I was in a position where I couldn’t afford to *not* get a good job)
anyways the day I gave up & changed it to Two Sigma (don’t judge me) Citadel gave me an interview lol
Threw together an interactive plot to explore the effect a stop-loss policy has on the distribution of returns based on the
@hudson_thames
reading group:
1. Shift in expected return is minimal except for extreme stop losses.
2. But you still avoid catastrophic tail events.
coding is my passion ❤️. I’m currently learning PEPE8 style guide to write clean code, and the best way to learn is to practice. do not be scared of failures 😤
Day 11 out of 30 of studying regressions *only*
(yes I’m in that phase of going through old materials in hope to find some novel insights and pass them off as my own)
not sure if it’s the best way but what I did was:
- look through other ppls projects/kaggle sols
- open an empty jupyter notebook and *type* the project line by line, and run the cell after each line so you know what’s going on
after doing this ~20 times you’ll be mid at coding
Friday quant fund janitor intern interview question - Linear Regression department (£300k base, 2 days wfh, protected nap time after lunch):
what’s the solution for OLS, if you enforce constraints where all coefficients are the same?
gappy on success:
“I think obsession helps. The world belongs to the obsessed…”
not to sound sigma grindset-y but if you’re not obsessed, someone else who is will lap you. if you’re not the smartest you have to give the universe a reason to grant your wishes; by being obsessed
yeah that looks hard but I bet this guy has never had to deal with the problem of underestimation of portfolio risk when the portfolio lies in the null space of your factor loadings on a Tuesday night
once I saw a senior quant missed his aim on the urinal and he panicked then ended up pissing all over himself so what makes you think quants would make good military commanders
it was rumoured that Van Gogh knew that combining Ridge Regression *and* PCR is superior than just using either one of it. but he couldn’t derive the estimator risk under mean squared error. it literally drove him suicidal. truly tragic. he was way ahead of his time 🕰️
been going to the gym for the last 9 months in an attempt to live a healthier lifestyle and finally hit 10 pull-ups today. might start posting vague stuffs about how you’re only a man if you can do 10 pull ups and make 7 figs if not your girlfriend will leave you stay tuned
I’ve never watched Margin Call. or The Big Short. or Billions. or Wolf of Wall Street. fwiw in my exp the ppl who always talk about these movies non-stop are the annoying-cringe-ngmi-hardo type 🤷♀️
sometimes I get DMs from ppl w a few yrs work exp asking "is it too late to pivot to quant/trading?". tbh I'm unqualified to give life advice, but idk given that you'll be working for the next ~30-40yrs, isn't it a bit sad to give up & limit yourself to what you can/can't do now?
a lot of ppl have asked about my opinion on Advances in Financial Machine Learning by Marcos Lopez de Prado. I have it there bcs sometimes I look at my own code and I hate myself for writing such shitty code but then I read the book and immediately feel better 🥰
few months back I was looking for a seat so I talked to many diff HF pods doing many diff things so I made flash cards to prepare for IVs. it was the most miserable moment in my life and after I was done I never ever looked at the cards again
dont ask about the choice of topics.
imagine spending your day masturbating thinking about which field/uni is 0.0001% marginally better than others. ngmi for sure. maybe if this fella spends those time grinding leetcode or Ridge Regression instead they wouldn’t even need to worry about this ranking BS
it’s my birthday and just bought myself a £104 The Book of Mormon ticket for tonight. not that big of a deal but it’s my first time doing this & expressing gratitude can only be a good thing. today we take a break from Ridge Reg and enjoy ourselves. have a good weekend all🙏
not embarking on this journey officially yet, but just chilling going through some literature on Random Forest. this particular paper is good tbh - comprehensive and also rigorous
should I embark on a journey of studying tree models (Random Forest/XGB/LightGBM wtv) *only* for 30 days…… can’t say I’m excited about it tho. but I think I’m not excited bcs I don’t even know enough about it….
had to delete and rebuild my python environment. a truly challenging moment. my humble request to everyone is to please respect my privacy during these difficult times.
I Studied Regressions *Only* for 30 Days So You Don't Have To But You Have To Subscribe
Lowkey-Advanced Ridge Regression (Part I)
as usual giving access to the lucky ones who RT
stolen content from
@ryxcommar
@AgustinLebron3
@JessicaNutt96
@negativecnvx
I received a lot of DMs abt learning Ridge Regression™️ and I don’t rly like to gatekeep so full disclosure I love reading this wiki it has lots of resources, papers, even code(!!!), & some interesting analysis on Ridge. do RT so more ppl can benefit🙏
you can’t just import sklearn and expect to make money. not meant to say this bcs its proprietary but this is what HF/props do:
import Money
for i in range(9999999999):
Money.print()
Jane Street even created a new language to speed up their loop look it up it’s 100% true
I don’t blame this person at all for not knowing but I blame all the LARPers putting out content about 18361730 different models in a handwavy way that makes newbies think it’s all just fitting models. when they themselves don’t even know basic linalg like QR/eigendecomp PCA etc
new post dropped:
All my homies hate LASSO
(it's meant to be free so everyone who RT will get access 😃)
🔮how good is LASSO at recovering sparsity?
💡 lesson from a macro HF PM in building +150 forecasting models
🏰 (lack of) invariance of regularisation
featuring stolen
I applied to many trading roles. most of them have some sort of leetcode/hackerrank and I never passed a single one(!). then I had to find some hacky way to break in. not that fun. the IRL project route IMO has higher EV but the leetcode grind route has much lower variance.
if you're a cs undergrad you're probably better off ditching leetcode forever in favor of getting good at willing ideas into existence. literally minmax solving IRL problems. don't do fake work
the cost of writing software approaches 0 in the limit so position yourself for that
If I tell you all the important features, should you include all of them?
why statsmodels >>> sklearn
probs my first "original" technical article (not quoting anyone on fintwit - mainly bcs I couldn't find anything on it)
as usual giving access to some who RT - thanks all 🙏
back from vacation my friends. been reading about Random Forest, and damn there are *so many* interesting tricks to get the most out of RF. and all of it require hacking the underlying model. friends don’t let friends import sklearn
this account used to be one of my fav accounts - he goes through really complex papers in 20mins trying to get the main points across etc. now it’s mostly just news about AI. no hate towards the guy but damn the content creator pipeline seems to rly ruins your soul.
i interviewed with the head of quant in one of the big 4 MMs. he asked a bunch of qs that I had to pass on since I had no clue. at the end he asked “are you willing to learn?” I said hell yes. he let me through lol. on the other hand my iv exp w sell-side juniors has been the…
any good reading on rates RV strategies? preferably the ones that go deep into the details. preferably have Sharpe ~2-2.5 after tcost. ideally not capacity constrained, net DV01 neutral and negatively correlated with carry and momentum. code attached is a plus. thanks in advance!
rates trading intern interview question (Financial Hacking-esque):
if rates stay the same for the next 10 years, what trade(s) would you consider putting on? what are the pros and cons of each implementation? be detailed
right now my life is at a high point. but like everything else, this too will pass. i can get fired, get sick, or lose all my money tmr.
have to keep my self grounded always. easy to forget that when you’ve “made it” bcs ppl treat you better wtv. have a good bank holiday all 🙏
why are you, as a grown man, still doing regime detection and classification? have you ever thought about learning continuous distribution which is just a teeny step above discrete distribution?
my first NPC purchase. never had a pair of sunglasses in my life ever. thanks to everyone who subscribed, didn’t know it’s possible to make ppl part with their hard earned money by just writing trivial stuffs about robust portfolio optimisation
I understand it’s annoying to get mass DMs from unqualified ppl. but at the same time I was there once where I was DMing ppl for a seat that I wasn’t qualified for. and a kind soul gave some leads that jumpstarted my HF career. takeaway is it’s okay to be *slightly* unreasonable
@palash_i_am
@__paleologo
"sir i traded my grandma's pa and made 0.5% from timing SPY this year. I have not yet internship but please give chance"
"sir i have worked at *obscure tier 8 prop shop* or FUNDED TRADER PROGRAM, pls consider*
for LASSO simps:
when should you use LASSO? look at the pic.
what does the formula mean?
it means if you regress "good" features against the "bad" ones, the betas should be less than epsilon.
in plain English it means LASSO struggles when good and bad features are correlated
janitor intern interview question for the soon-to-be-launched Generalised Ridge Quant Fund: Linear Regression 101
what are the differences between the 3 statements? (v easy)
for each of the statement, what are the assumptions needed for the statement to be true? (medium)
once you realize the markets are rigged by big, rich hedge funds, you almost have no choice but to JOIN MY COURSE
where you will:
-start trading on day 1!!
-make 11.5% in 13 days!!
-trade crack spread as retail alongside the best commods pods!!
hedge funds HATE this one course
Pinned Strategy of The Week:
FX strat based on German rates
Method: Model FX rets in different term structure environments & punt🥳
Got the idea from
@donnelly_brent
's book. Half of my playbook is stealing ideas from discretionary traders & try to engineer systematic signals🙂
been receiving CVs from students who guessed my work email & *IMO*:
-some of you need more projects esp if no relevant exp
-some of you have decent projects but write it like you hate your projects
PS: I have 0 influence in hiring I myself struggle to keep my job :’) so PAYOR
once I was DMing portfolio managers my profile + work and received ok ish responses. after that whenever I DM I’d say “I donated 10bucks to xxx charity under your name to make it worth your time reading this”. the reply rate was much higher but it was defo expensive 😂
really sad reading about the history of this place. back then when quants wanted to get filled they had to fight their counterparty to death (also known as Fill-or-Kill order). even XTX was reported to lose around ~250 quants, now they decide to double down on GPUs