arb8020 @arb8020 Twitter profile | Pikagi

Pikagi

arb8020

@arb8020

3,765

Followers

623

Following

102

Media

2,968

Statuses

its all just a matter of time

Joined July 2022

Don't wanna be here? Send us removal request.

Pinned Tweet

@arb8020

arb8020

2 years

hey here's a super rough draft on my thoughts on how i would speedrun coding competency (basic python/pandas) for beginner quant stuff if you take away nothing else: embrace fuck around and find out, let what you find fun guide you skip around if you want, there's titles

9

63

376

Last Seen Profiles

@BnxLovekid

@gumi_goods293

@bokeplokalmalam

@MPEditsYT

@one_reamman

@KinorcKinorc

@comedian_mako

@marmur2010

@sheep_gardenie

@JethBodeen

@EmmaObuya_ke

@ConfessionsTxsu

@Catronicus

@melodicrhymez

@DAMIADENUGA

@teena_thach

@MoraFumero

@clareliedstrand

@oyabaremeigen

@CU_ethnolab

@stw_pdg

@mr_upskirt

@anpaure

@VerhaegTress

@TrineFlockhart

@harusame_udonn

@m0rkim00

@watermelonnanas

@AP_Athletics

@David9319574839

@lukasek_josef

@RT__feet20

@minabuttell

@_kyletaylor1

@NickZhurav

@jrk74831595

@arb8020

arb8020

24 days

just finished up a quant trading internship this summer, turns out i don't really want to be a trader full time still really grateful for it! taught me i want more work where i have some kind of output (engineering, research)

17

5

357

@arb8020

arb8020

2 years

hello: this is my 'quant' analysis of @worstcontrarian 's idea here: he hypothesizes that we can fade 'moon' candles, or bigger than normal price movements, and make money on average doing so

16

39

253

@arb8020

arb8020

2 years

if you're interested in microstructure, orderbooks, and coding i think a toy orderbook project is fun and instructive given an initial snapshot of an orderbook, and all updates to price levels, maintain a copy of the true state of the orderbook

11

15

239

@arb8020

arb8020

2 years

hello: here is my quant investigation into the following tweet from @hyuktrades on average, it seems like newstrading is pretty profitable if you can get the news fast, and execute in the right direction but hyuk says it's competitive: how competitve?

14

28

206

@arb8020

arb8020

2 years

statquest and 3blue1brown are my favorite resources for making probability/stats/ml/math more intuitive here's an organized list of links of videos to watch that will take you from 0 to understanding essential concepts for quant research

Tweet card media

QR Youtube Watch List | Notion

Probability/Statistics: Statquest

work-in-progress.notion.site

7

24

158

@arb8020

arb8020

2 years

is it just me or is the answer for data analysis stuff always fast/slow emas linear/regularized regressions rank/zscore ensemble/average a bunch of shit

8

9

155

@arb8020

arb8020

2 years

import pandas as np import numpy as pd

9

6

152

@arb8020

arb8020

2 years

trading has ruined me i am trying to bet on the super bowl for fun w my friends and i cant stop thinking about how i have no edge and shouldn't bet at all

8

0

134

@arb8020

arb8020

2 years

just finished reading winning sports betting by masaru kanemoto really good book, lots of trading parallels, gives a good high level overview of how to approach sports betting if you are serious about winning

5

7

122

@arb8020

arb8020

2 years

is there any edge in trading options around stock splits? here's some initial data analysis me and my friends did into that idea:

3

13

109

@arb8020

arb8020

2 years

correct me if wrong but trades executed probably give more information than book/resting liquidity ? since to trade you have to cross the spread it likely signals more information/ stronger belief in higher future price over the next N ticks than just liquidity on bid/ask right

15

7

106

@arb8020

arb8020

3 months

it has come to my attention that, despite being nerd sniped by ML, quantitative trading is really fucking fun

3

2

108

@arb8020

arb8020

2 years

looking at the first 10 questions of the 150 most frequently asked quant interview questions and i already don't know what some of these words mean Ornstein-Uhlenbeck SDE?? what .

17

3

75

@arb8020

arb8020

2 years

if polars is a faster pandas why is it not called pandas express

3

2

70

@arb8020

arb8020

1 year

would there be any interest in reading explanations of solutions to common interview questions? im working on my ability to explain things + problem solve so id be happy to do that in public

10

2

69

@arb8020

arb8020

2 years

from now on i will just start running a strategy in prod the loss of money will be incentive to improve it as fast as possible before it loses more money fuck it, we ball

@therobotjames

Robot James 🤖🏖

2 years

Generally, backtesting trading strategies sucks. The best case is something so big and obvious that the accuracy of the sim doesn't matter much. So, the knowingly "finger-in-the-air' sim is cool. Cos, as soon as it matters, things get very expensive in data qty and processing.

24

11

175

10

1

67

@arb8020

arb8020

2 years

im trying to do options things and ive decided i should prob read sinclair's books first

11

2

65

@arb8020

arb8020

2 years

every day i open my research notebook and commit crimes against good coding practices

2

1

64

@arb8020

arb8020

24 days

@AgustinLebron3 for sure! i think the reflection is still in progress but it was largely due to the lack of a tangible output. i still think that trading teaches a lot of valuable meta skills but i miss being able to point to something and saying 'i built this' or 'i discovered this'

5

0

65

@arb8020

arb8020

2 years

data visualization is my passion see if u can spot what i did wrong

Tweet media one

16

1

62

@arb8020

arb8020

1 year

just got exploited for being a price insensitive buyer for coffee at the airport gate i don’t wanna talk about how much i paid

8

0

62

@arb8020

arb8020

8 months

@MikeDugar peter carroll

0

0

52

@arb8020

arb8020

2 years

this is one of my favorite videos of all time on youtube presents two very key mental models that are core to who i am and how i think

Tweet card media

Think deeply about simple things

Main site: http://www.misterwootube.com Second channel (for teachers): http://www.youtube.com/misterwootube2Connect with me on Twitter (http://twitter.com/mi...

www.youtube.com

1

7

49

@arb8020

arb8020

2 years

some initial thoughts on pairs trading i don't know exactly how it works, but here's some of my intuition/wild guesses on it right now: will follow this up with some code and data later this week

5

6

50

@arb8020

arb8020

2 years

poker tourney in 4 hours gimme the pareto principle things to know (20% of info -> 80% of results)

17

2

49

@arb8020

arb8020

2 years

i cant say for certain if i have edge click trading futures but i can say that i've learned a lot about markets from it best trading book is the orderbook: go read it

4

3

49

@arb8020

arb8020

2 years

what makes citadel/citsec so good? can't just be PFOF since virtu has that as well and doesn't have the same performance thoughts?

10

1

50

@arb8020

arb8020

2 years

missed an A on my abstract algebra class by 0.06% didn't do anything all semester and only studied for the last two weeks of school i have found nash equilibrium

4

0

44

@arb8020

arb8020

2 years

@quant_arb

Tweet media one

2

1

42

@arb8020

arb8020

2 years

super useful thread! give this a read, important takeaway at the end

@macrocephalopod

cephalopod

@macrocephalopod

3 years

Here's a quick primer on how to build a quant reversion strategy. First step is to find a price series that looks like it may have some reverting behaviour -- this is a random stock I found that looks reasonable but you could use an ETF, currency, whatever really.

Tweet media one

17

69

460

1

3

46

@arb8020

arb8020

2 years

i know im very late to this but book pressure plots look so cute idk what it is just v satisfying to look at

Tweet media one

1

0

44

@arb8020

arb8020

2 years

java, coffee: good java, language: bad

1

0

39

@arb8020

arb8020

2 years

colin galen's recent series of videos on problem solving and etc are very good, would recommend watching them lays out strong frameworks for thinking about problems and learning new concepts here's a link to one of them, check out the other ones!

Tweet card media

Focusing Your Unconscious Mind: Learn Hard Concepts Intuitively (And...

A general learning method for learning and understanding hard concepts intuitively/deeply/obviously, and for long periods - up to years at a time.I’ve put in...

www.youtube.com

3

2

40

@arb8020

arb8020

2 years

hope you all enjoyed reading this, let me know if you have any feedback, ideas on extending this analysis, or questions for me!

9

0

40

@arb8020

arb8020

2 years

project idea for people who want to start getting their hands dirty w basic quant stuff should be reasonably simple to do this - i can thread abt general guidelines/steps but i’d argue to some extent figuring it out yourself is helpful

@mattyklein_

Matt Klein

2 years

anyone backtested magic internet money on days where the burger stock market is closed

3

0

20

2

1

37

@arb8020

arb8020

2 years

for basic probability stuff, this link from jane street is an incredible starting point: doesn't rly go over distributions as much as i'd like but i think statquest should have you covered on that front

1

2

33

@arb8020

arb8020

2 years

good data to practice on is 's free incremental L2 data for the first of the month

Tweet card media

The most granular data for cryptocurrency markets — Tardis.dev

Historical tick-level order book data, trades, funding, liquidations, options chains and more. API access and downloadable CSV files. Real-time consolidated market data streaming API via client...

1

1

32

@arb8020

arb8020

2 years

sounds a lot to me like stuff i’ve been hearing about pairs trading trading costs and are frequently too high to put on both legs: just figure out which leg to bet

Tweet media one

1

1

32

@arb8020

arb8020

2 years

not trading related but may take some time to work on the NFL big data bowl if anyone's interested in that i can publish what we did/thought process after the 9th

3

0

32

@arb8020

arb8020

1 year

this is cool

1

3

31

@arb8020

arb8020

1 year

good devs keep telling me the best way to get better at coding is contributing to open source - helps onboard into new codebases faster - familiarization with good code style/ documentation - working with real stuff i’m already sold but would love to hear about more benefits!

5

0

29

@arb8020

arb8020

2 years

ty @therobotjames for explaining stuff to me and ty to @worstcontrarian for the idea now i have to do my homework

6

0

32

@arb8020

arb8020

2 years

born to trade forced to do math

1

1

28

@arb8020

arb8020

2 years

@thiccythot_ why did u blur out ur face when ur face is in ur pfp

1

0

29

@arb8020

arb8020

2 years

probably gonna rewatch some relevant statquest today and take notes to get the fundamentals locked in before i start rereading ESL can publish notes on either if interest is there

4

3

28

@arb8020

arb8020

1 year

NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO NEW MATH VIDEO

Tweet card media

But what is the Central Limit Theorem?

A visual introduction to probability's most important theoremHelp fund future projects: https://www.patreon.com/3blue1brownSpecial thanks to these lovely sup...

www.youtube.com

2

0

26

@arb8020

arb8020

2 years

this thanksgiving, i learned that i've been missing out on the most important aspect of gratitude practice here's what it is, and how to implement it! all credit goes to @hubermanlab , check out the video for tons more detail on the science of gratitude

1

3

29

@arb8020

arb8020

1 year

i keep overhearing people saying ‘who’s gonna carry the boats??’ at the gym very based but i started laughing and almost failed my bench rep

2

0

29

@arb8020

arb8020

2 years

guys here’s the potluck me and my friends had 🥺🥺🥺

Tweet media one

5

0

29

@arb8020

arb8020

2 years

this blog will guide you better than i ever could on the statistical competency side: recommends chapters of the famous ESL book and a hands-on google ML course for practicality (linked)

1

1

28

@arb8020

arb8020

8 months

what i learned in 2023 is that the stupid grindset chase your dreams stuff actually works just gotta fuck around and find out happy holidays

2

1

28

@arb8020

arb8020

2 years

here’s a cool hack to make sure you don’t overfit your linear models! print out the data and draw the line by hand 👍

2

1

28

@arb8020

arb8020

2 years

it is so easy to fall off the track that good habits have built you i took some rest from friday -> today (arguably too much) and momentum is shot need to get the momentum back in my favor and keep grinding

3

2

25

@arb8020

arb8020

1 year

tag yourself im fuck it we ball

Tweet media one

11

0

27

@arb8020

arb8020

2 years

kaggle courses: (do in order) very hands on learning: makes you actually code i chose kaggle bc its where i learned pandas, + from friends the other two intro to programming/python links are great as well

Learn Pandas Tutorials

Solve short hands-on challenges to perfect your data manipulation skills.

1

0

27

@arb8020

arb8020

2 years

i’ve peaked

Tweet media one

Tweet media two

2

0

27

@arb8020

arb8020

2 years

LATEX ON SUBSTACK LETS GOOO maybe i’ll actually use my substack

0

1

27

@arb8020

arb8020

2 years

for 2023 i want to be someone who reflects more i'm using google calendar as my journal: here's how

2

1

26

@arb8020

arb8020

2 years

@OrthogonalAlpha just a guess sampling, features, fitting intuition is that priorities should be: start with good data (sampling), then how you interpret that data (features), then how you extract insight from that data (fitting)

0

1

26

@arb8020

arb8020

2 years

always wanna start w an understanding of the underlying market effect here: basically we think that 'moon' candles represent aggressive buyers/sellers who are willing to slip through the book to execute immediately since fair value probably hasn't changed, we fade the movement

1

0

25

@arb8020

arb8020

1 year

battle of the century: dehydration vs frequent peeing

5

0

26

@arb8020

arb8020

2 years

let's look at the CFTC case on Navinder Singh Sarao, commonly referred to as Nav while what he did was illegal, its still pretty interesting and there's a lot to be learned here's (sort of) a visual of what he was doing

2

6

26

@arb8020

arb8020

1 year

i get confused when people cross the street early after they see me do it i calculated my survival chances not yours

3

0

25

@arb8020

arb8020

2 years

ethan chlebowski is my favorite cooking youtuber rn dives into 'why' behind the things you do when you cook - similar to kenji lopez alt, adam ragusea, salt fat acid heat, etc great video, i cooked my chicken breast like this today

Tweet card media

Perfect Chicken Breast at home (According to Food Science)

For a limited time only, get up to $100 off your first 4 boxes! Follow my link https://bit.ly/BrightCellarsChlebowski5 to take the taste palate quiz to see y...

www.youtube.com

3

1

25

@arb8020

arb8020

2 years

ESL is more of a reference guide: Intro to Statistical Learning is better for learning if you like books personally i'd look up the concepts covered and see if StatQuest has videos on it: always my go to for strong fundamental/intuitive understanding

Tweet card media

The Main Ideas of Fitting a Line to Data (The Main Ideas of Least...

Fitting a line to data is actually pretty straightforward.For a complete index of all the StatQuest videos, check out:https://statquest.org/video-index/If yo...

www.youtube.com

1

0

25

@arb8020

arb8020

1 year

@quant_arb smh real ones know hummingbot is great if you use the avellanada and stoikov settings it fixes all the issues and makes it v profitable and unexploitable

4

0

25

@arb8020

arb8020

2 years

so sports betting books can straight up ban you for playing certain strategies? wild

5

2

23

@arb8020

arb8020

2 years

analysis done on 1min OHLCV binance BTC/USDT data assumed fills at closes of candle, didn't account for slippage, etc mean log return at each period may have been the wrong way to interpret the events didn't look at how this compares to buy/hold, etc didn't look at sharpe ratio

3

0

23

@arb8020

arb8020

2 years

didn’t get into the firm i was supposed to hear back from we keep building

5

0

21

@arb8020

arb8020

2 years

it is time to teach myself abstract algebra in 2 weeks planning to: memorize some key definitions/proofs do as many practice problems as i can do my hw/midterms again do my practice final ace final pls let me know if i should add to, or remove from that strategy

6

0

23

@arb8020

arb8020

2 years

writing up thread from max dama q/a hopefully should be done tomorrow latest

1

0

21

@arb8020

arb8020

2 years

INTRO before getting into resources: i'd like to emphasize that you must actually do things to get good practical things like reading documentation to learn a library, debugging code, engineering data, are best learned by doing it. trust.

1

0

23

@arb8020

arb8020

2 years

@thodoha @nope_its_lily this one ?

Tweet card media

GitHub - thelilypad/orderbook_simulator: A simple, extensible implementation of the limit order...

A simple, extensible implementation of the limit order book. - thelilypad/orderbook_simulator

0

2

23

@arb8020

arb8020

24 days

@AgustinLebron3 i think in a role thats more skewed towards research or dev i might be happier, but at least at the firm i was at i think most of the emphasis was on building skills to be good at discretionary trading which was less tangible

2

0

23

@arb8020

arb8020

2 years

@therobotjames this is true very +EV to know this

1

0

22

@arb8020

arb8020

2 years

numerical analysis actually seems super useful too bad we did only one coding project and the rest of the hw was direct calculations oh and the final is all theoretical, where we've had little to no practice might just try to retake this class w a professor i understand better

3

0

20

@arb8020

arb8020

2 years

i’ve been purposefully avoiding binging content recently and i actually really like it more satisfying to read a chapter per day and give myself time to think about and get excited for the next chapter rather than immediately reading ahead same for tv shows, etc

2

0

21

@arb8020

arb8020

2 years

taking a forecasting class at school and i really need jpow to hike interest rates by 25bps since my confidence interval i submitted is 4.75 @ 4.75 🙏

1

0

21

@arb8020

arb8020

2 years

we define a moon candle as one that is bigger than normal candle size: (candle high - candle low) 'bigger than normal' means that it is some amount of standard deviations greater than the average candle size over some recent period (i did over the past 60min)

1

0

22

@arb8020

arb8020

1 year

@KrisAbdelmessih

Building a Futures Market Making Bot - Winning the UChicago Trading Competition, Part 1

How we built a customizable trading bot and won the Futures Market Making (MM) case.

2

3

22

@arb8020

arb8020

2 years

@quant_arb you really just need regression + regularization model selection (bias/variance, etc) ensembling i think there’s more value in deploying something that mostly works and watching markouts, trades, etc than learning a lot of ML stuff you might not need

1

0

21

@arb8020

arb8020

2 years

guys i’m attending a Q+A with max dama tonight any questions you want answered?

1

0

21

@arb8020

arb8020

5 months

just finished "performance analysis and tuning on modern CPUs" by Denis Bakhvalov, was a great read! trying to figure out the best way to arrange my notes for my own understanding + get more practice im thinking working through some open ended examples but not sure

3

0

20

@arb8020

arb8020

2 years

im thinking i build a probability distribution based on option implied pricing, then i can lift offers on kalshi yes/no market if we disagree by more than cost to trade then we'd adjust position as new option implied information comes in?

3

0

18

@arb8020

arb8020

2 years

dont tell anyone but i think i got hooked on trading when i made 25 bucks off an AMD put the day i turned 18

2

0

20

@arb8020

arb8020

2 years

we're getting paid for providing liquidity to the price insensitive player, as well as for helping converge the asset price back to fair value the risk we take is that fair value has changed, and we're selling too cheap/buying too expensive

1

0

20

@arb8020

arb8020

1 year

happy birthday @KrisAbdelmessih you need to follow this man and read his content he’s got an incredible wiki of trading stuff and just generally good advice and thoughts on things i owe him a ton

5

0

21

@arb8020

arb8020

2 years

if im price insensitive: i just want to buy a ton of the asset, i might not give a fuck and lift whatever i need, say 69 (nice) lots ill fill 8 at 101, 27 at 102, and 34 at 103 so i pay too much because i demand liquidity now, and i rip the market higher

1

0

21

@arb8020

arb8020

2 years

so a candle is 'big' if: candlesize > avg_size + stdv_size * (sigma) where sigma is our threshold for how many standard deviations above the mean the candle has to be to be statistically 'big' (i chose sigma=5)

1

0

21

@arb8020

arb8020

2 years

im gonna cumulatively sum the mean log return for each of our periods instead, because we should get graphs that are easier to read than whatever that mess was looks like we were right: log returns on average trend in the opposite direction of our moon candle

Tweet media one

Tweet media two

3

0

20

@arb8020

arb8020

1 year

@SpencerBosma trade the opposite direction of ur signal

2

0

20

@arb8020

arb8020

2 years

we think fair value of the asset is probably still between 100:101, unless the person who just lifted 69 lots knows something and the fair value of the asset changed so on average we think we can fade this movement and expect price to return to 100 @101

1

2

20

@arb8020

arb8020

1 year

many people ask me: if pandas is so good, why isn’t there a pandas 2? well folks, there is

2

1

20

@arb8020

arb8020

5 months

got sucked into performance optimization for CPUs these past few days thanks to rewatching this video here's the journey that i took, and the links i got the most value out of!

Tweet card media

CppCon 2017: Carl Cook “When a Microsecond Is an Eternity: High...

http://CppCon.org—Presentation Slides, PDFs, Source Code and other presenter materials are available at: https://github.com/CppCon/CppCon2017—Automated tradi...

www.youtube.com

2

2

19

@arb8020

arb8020

2 years

heres a high level overview video of most of what u need from data structure/algorithms, p funny too cheatsheet on interview-level data structures and algorithms: try to understand the info, don't worry too much abt the questions

Tweet card media

Data structures and algorithms study cheatsheets for coding interviews | Tech Interview Handbook

Study guides for coding interviews with focus on data structures and algorithms, including practice questions, techniques, time complexity and recommended resources

www.techinterviewhandbook.org

1

0

20

@arb8020

arb8020

1 year

i want to mute the term AI so bad

2

0

20

@arb8020

arb8020

1 year

wdyt of this CEX i found

Tweet media one

3

0

20

@arb8020

arb8020

2 years

im a depth of market guy, so i'll explain the microstructure here, can skip to data say your asset's orderbook looks like the below, price in the middle, bids on left, asks on right 105x55 104x40 103x34 102x27 101x8 11x100 23x99 35x98 44x97 63x96

1

0

20

@arb8020

arb8020

1 year

statquest has the most intuitive and easy to follow explanations for statistics concepts on youtube genuinely criminal that @joshuastarmer doesn't have 1mil subscribers yet on youtube

2

1

20