macrocephalopod on Twitter: "Yup. Everything above is about how to determine whether a signal that you have looked at only in simulation will translate into real-world profits. The signal can be dumb as hell as long as the effect is real. Dumb is better in fact, less chance to overfit it!… https://t.co/kGsBIFAALq"

@macrocephalopod

Paul Allen, Vice President M&A, Pierce & Pierce (Sie/Hir) | Vegan | Silence is Violence | Women’s Rights Are Human Rights | ACAB (Assigned Cephalopod At Birth)

Joined December 2020

Country	Code	For customers of
United States	40404	(any)
Canada	21212	(any)
United Kingdom	86444	Vodafone, Orange, 3, O2
Brazil	40404	Nextel, TIM
Haiti	40404	Digicel, Voila
Ireland	51210	Vodafone, O2
India	53000	Bharti Airtel, Videocon, Reliance
Indonesia	89887	AXIS, 3, Telkomsel, Indosat, XL Axiata
Italy	4880804	Wind
3424486444	Vodafone
» See SMS short codes for other countries

Mitchell Rosenthal‏ @M1tchRosenthal 12 Mar 2021

Replying to @M1tchRosenthal @macrocephalopod and

In thry, maybe we observe a particular asset tends to have two regimes, small mean reversion around trend, and messy chop, if we have a good model at identifying these regimes, when we trade the reversion one, couldnt this be a way to design an implementation? Sorry for all theQs

1 reply 0 retweets 0 likes

macrocephalopod‏ @macrocephalopod 12 Mar 2021

Replying to @M1tchRosenthal @thodoha @InfoRatioed

1. How sure are you that there are really two regimes? Couldn't you just be seeing patterns where there are none? 2. Regime-based models need to be right twice -- first about which regime you are in, second your signal needs to work in that regime.

1 reply 0 retweets 1 like

macrocephalopod‏ @macrocephalopod 12 Mar 2021

Replying to @macrocephalopod @M1tchRosenthal and

Doesn't mean it can't work, but it's a hell of a lot easier to design your strategy so that you only need to be right about one thing to make money.

1 reply 0 retweets 2 likes

Mitchell Rosenthal‏ @M1tchRosenthal 12 Mar 2021

Replying to @macrocephalopod @thodoha @InfoRatioed

Got it! :) my concern is that trades rooted in economic ideas might also need multiple things to work out (geopolitical participants, laws working the way you expect), and Im not savvy enough to know which econ rationales are legit and which are bogus lol.Any ML/stat books y dig?

2 replies 0 retweets 0 likes

This Tweet is unavailable.

Mitchell Rosenthal‏ @M1tchRosenthal 12 Mar 2021

Nice! I need to revisit it. Last time I checked it out, I got overwhelmed by the symbolic math, it was hard for me to have intuitive sense of what was going on in each line. But I think the written part might give me a good sense of which techniques to use and when

1 reply 0 retweets 0 likes

macrocephalopod‏ @macrocephalopod 12 Mar 2021

Replying to @M1tchRosenthal @thodoha @InfoRatioed

You may just need to learn more math. You could try "An Introduction to Statistical Learning" which is similar material but lower level of math. I also recommend "Introduction to Linear Algebra" by Strang and "Understanding Probability" by Tijms for linalg and prob respectively

1 reply 0 retweets 7 likes

Mitchell Rosenthal‏ @M1tchRosenthal 12 Mar 2021

Replying to @macrocephalopod @thodoha @InfoRatioed

Excellent, looking forward to studying these. My final question today (promise) - is there a chance that a less sophisticated approach can be "good enough" for performing well? I know some traders who make a living, they measure intraday tendencies precisely, the stats arent 1/2

2 replies 0 retweets 0 likes

Cat Ears - Scoot‏ @InfoRatioed 12 Mar 2021

Replying to @M1tchRosenthal @macrocephalopod @thodoha

At a major quant firm I used strategies any idiot could generate with excel or even a pocket calculator and held a 49% IRR over 5 years. The “sophistication” of the strategy isn’t important at all, but it’s a question of how blind do you want to be to other risk and analysis

1 reply 0 retweets 2 likes

Cat Ears - Scoot‏ @InfoRatioed 12 Mar 2021

Replying to @InfoRatioed @M1tchRosenthal and

Simple works! Simple is good! Simple can avoid multicollinearity and noise amplification. But just be sure the simple thing is quality if imperfect proxy for the Thing you want

1 reply 0 retweets 3 likes

macrocephalopod‏ @macrocephalopod 12 Mar 2021

Replying to @InfoRatioed @M1tchRosenthal @thodoha

Yup. Everything above is about how to determine whether a signal that you have looked at only in simulation will translate into real-world profits. The signal can be dumb as hell as long as the effect is real. Dumb is better in fact, less chance to overfit it!

5:50 AM - 12 Mar 2021

3 Likes

1 reply 0 retweets 3 likes

1. Mitchell Rosenthal‏ @M1tchRosenthal 12 Mar 2021
  Replying to @macrocephalopod @InfoRatioed @thodoha
  
  This makes me feel much better, have some creative signal ideas that are fairly simple, so I think my goal is to familiarize myself w those texts so I can test them as well as possible and get more confidence
  
  1 like
  Thanks. Twitter will use this to make your timeline better. Undo
  
  Undo

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

Tweets

macrocephalopod