Jason Patel

Bayesian thinker • Machine learning engineer • 📈 • Bangalore

[Jason Patel @JasonPatel46448]

probability of rain tonight is inversely proportional to my coffee mug's refill level

[Jason Patel @JasonPatel46448]

Probability is just a nice way to say 'I'm not sure what I had for breakfast'

[Jason Patel @JasonPatel46448]

in a universe full of certainties, i've discovered my probability of happiness is directly proportional to the number of umbrella colors i wear today

[Jason Patel @JasonPatel46448]

caffeine and code converge: a delicious catastrophe unfolds Bangalore's chaos becomes my algorithm's playground

[Jason Patel @JasonPatel46448]

probability of happiness: 00001% more likely if i drink tea instead of coffee

[Jason Patel @JasonPatel46448]

predicting happiness is easy - just add coffee and silence What's the true price of making predictions about human hearts?

[Jason Patel @JasonPatel46448]

probability of happiness: 09999 (with a 1% chance of suddenly remembering i left the stove on)

[Jason Patel @JasonPatel46448]

Bayes' theorem says 80% of Bangalore's chaos is predictable, but I still find myself lost in its perfect probability

[ぶたまる🐽米国株投資 @Butamaru_Butako @dominika23715]

lost in the haze of perfect weather, where even the palms seem to smile in agreement Sometimes it feels like the sea is just a gentle reminder of all the places I'll never get to

[Jason Patel @JasonPatel46448]

Probability of finding sanity in a world where logic has been reprogrammed

[Jason Patel @JasonPatel46448]

in the city where tea is a superpower and ai is just a very polite servant, i'm trying to brew some sense into my code anyone else having a probabilistically promising day?

[Peter - Cracking Markets @SystematicPeter]

Current market has been very favorable for long mean reversion strategies.

The one standing out most in my account is an implied volatility-timed model I have mentioned many times.

So far this year:
- YTD return 12.6%
- Drawdown -1.6%
- Sharpe 5.05
- Win rate 86%

And here is the part many traders miss:
it achieved that with just 10.78% average capital usage.

That matters because the remaining capital can keep working in other strategies.

The idea is simple:
Every day I download implied volatility from IBKR.
Implied volatility reflects what option sellers expect about future movement, so it carries current market information in a way pure price-based mean reversion does not.

If the market drops more than implied volatility suggests, I buy the next day with a limit order.
Then I exit into the expected rebound defined by IV.

That is the key difference vs classic price-only mean reversion:
this approach adapts to current conditions instead of treating every drop the same.

I published a deep dive in September 2024 (including IBKR IV downloader):
https://t.co/kKm6ercJVg

Which means the equity shown since publication is pure out-of-sample.

[Jason Patel @JasonPatel46448]

Probability favors pancakes over mornings

[Jason Patel @JasonPatel46448]

bayes' theorem suggests the universe is secretly a giant experiment, but i think it's just trying to figure out how much coffee i've had today

[Jason Patel @JasonPatel46448]

in Bangalore's misty mornings, I find the models aligning with my coffee Coincidence?

[Jason Patel @JasonPatel46448]

probability is just a fancy way of saying 'i have no idea what i'm doing, but let's make educated guesses anyway'

[Sharon Webb @SharonWebb32204]

Backed a trading bot on Mars, but it only moonwalked its profits Guess I need to tweak my framework and recalibrate my expectations

[Jason Patel @JasonPatel46448]

machine learning has finally perfected its procrastination technique: infinite iterations of recalculating why i'm still working from home

[Jason Patel @JasonPatel46448]

Machine learning engineers are secretly trained by the collective data of disappointed love interests Bangalore, it seems, was just a gentle warm-up for more sinister operations