dan_is_coding's profile picture. Coding.  HFT.  Crypto.   Rust.

Daniel Cook

@dan_is_coding

Coding. HFT. Crypto. Rust.

Indirect answer: Make your FV the price at which, given a fill and the subsequent hedge, the EV of your realized PNL = 0. So it should depend on your risk holding time expectation conditional on a fill at each moment.

here i have the same feature, but running on 1m candles and 1s candles respectively. each produces their own fair value, but work on different turnover periods. these fair values can end up being wildly different under vol (1%+). how can i normalize their fair values?

BeatzXBT's tweet image. here i have the same feature, but running on 1m candles and 1s candles respectively.

each produces their own fair value, but work on different turnover periods. these fair values can end up being wildly different under vol (1%+). 

how can i normalize their fair values?
BeatzXBT's tweet image. here i have the same feature, but running on 1m candles and 1s candles respectively.

each produces their own fair value, but work on different turnover periods. these fair values can end up being wildly different under vol (1%+). 

how can i normalize their fair values?


New LinkedIn is wild..

dan_is_coding's tweet image. New LinkedIn is wild..

Daniel Cook รีโพสต์แล้ว
therobotjames's tweet image.

Pray for this to be your competition…

In my previous role, I used `tokio-console` to optimize the HFT system. It helped me cut latency from 2 sec to 1-100 millisec by identifying bottlenecks. Occasionally, adding a small delay in a channel improved performance, especially with multiple data sources flooding a channel

wiseaidev's tweet image. In my previous role, I used `tokio-console` to optimize the HFT system. It helped me cut latency from 2 sec to 1-100 millisec by identifying bottlenecks. Occasionally, adding a small delay in a channel improved performance, especially with multiple data sources flooding a channel


The traders live-adjusting the HFT algo parameters

AI has reached its potential



Expanding out the latency curve from HFT to MF. Looking for resources/experts on data engineering and platform design in mid-frequency- ingestion, cleaning, storage, replay, of a wide variety of datasets. What is the elegant API sweet spot of a midfreq strat?


Every Rust developer should read: doc.rust-lang.org/nomicon/ The (dumb) gatekeeping in the intro obscures what is certainly the best guide to using Rust effectively.

To those who want to learn more about Rust's memory management model (& other interesting topics) It's an integral part of knowing what's happening under the hood and no one does it better imo youtube.com/@coredumpped?s…



Systematic midfreq executing through fpga and microwave

We need a slur for beginners in a hobby who already bought the most expensive gear



“Searching for alphas” - no alpha. “Wtf is this free money? Does no one else even look at this s**t?” - lots of alpha

my PM effectively said "you're too dumb to find good trades stick to portfolio optimisation instead then you'll probably be useful" 🤷‍♀️



Friend is a solid hire. She ships.

my friend wrote a web app for her masters thesis in R. it’s literally one file with BOTH ui AND server components. i’m (react dev) in shambles p.s. she allowed me to post it

enjojoyy's tweet image. my friend wrote a web app for her masters thesis in R. it’s literally one file with BOTH ui AND server components. 

i’m (react dev) in shambles 

p.s. she allowed me to post it
enjojoyy's tweet image. my friend wrote a web app for her masters thesis in R. it’s literally one file with BOTH ui AND server components. 

i’m (react dev) in shambles 

p.s. she allowed me to post it


Hate speech

How to ruin a data scientist:

svpino's tweet image. How to ruin a data scientist:


Daniel Cook รีโพสต์แล้ว

There is no need to mention Rust because 50 people will do it in the comments.


Holy crap SO MUCH this

__femb0t's tweet image.


This tweet slaps.

my quoting system is finally working like I wanted, this is actually a good representation of the logic behind it:

gumby_arc's tweet image. my quoting system is finally working like I wanted, this is actually a good representation of the logic behind it:


My IOCs going through a shitty crypto matching engine to hit the orderbook

Bro forgot to overthink



You’re paying $1m for the culture. After 12, kids’ main influence is peers. So you’re paying for the kids main influence to be families of winners who can (and want to) pay $1m for their kids. Not saying it’s always right, or that it always works. But that’s why you pay. It’s…

if you spend $15k/yr for private schooling (low) + $150k additional for private college that amounts to $1m of money you could’ve given them at 30 that has probably negligibly improved their life you need to point to some pretty tangible life improvements for it to be worth

Duderichy's tweet image. if you spend $15k/yr for private schooling (low) + $150k additional for private college that amounts to $1m of money you could’ve given them at 30 that has probably negligibly improved their life

you need to point to some pretty tangible life improvements for it to be worth


Time to add drift to the todo list…

just submitted a $13 SOL market order on drift and got filled at $482 per SOL



GPT> Generate a simple diagram of a standard matching engine design The diagram: 🤯

dan_is_coding's tweet image. GPT> Generate a simple diagram of a standard matching engine design

The diagram: 🤯

Loading...

Something went wrong.


Something went wrong.