class @classpath

Quant developer Crypto trading enthusiast Joined March 2016

Tweets

85
Followers

283
Following

94
Likes

6K

class @classpath

4 months ago

Let me get this straight.. Binance crashes the price of LAYER, leading to a -$45M loss for them since they own 15% of the supply. Their manipulation causes ~500k in liquidations on 5/7/2025, leading to a 5-15k gain in liquidation fees? Am i missing something?

Tweazyy @Tweazyk

4 months ago

1 0 7 878 0

0 0 1 139 1

class @classpath

4 months ago

0 0 7 99 0

Download Gif

class @classpath

7 months ago

Ran a simple memory access test for my OS class. The jump at ~8 pages reveals the TLB cache size—just a couple of loops and you can measure it yourself! 📊 Practically, a TLB miss makes memory access 5x slower.

0 0 1 92 0

Download Image

class @classpath

10 months ago

I doubted this claim so I priced the second contract (btc hits 100k in Nov) with a monte carlo. Polymarket: 40% for YES My model: 40.7% for YES

otw2 @imotw2

10 months ago

I doubted this claim so I priced the second contract (btc hits 100k in Nov) with a monte carlo. Polymarket: 40% for YES My model: 40.7% for YES

3 1 15 6K 4

Download Image

2 0 7 2K 4

class @classpath

12 months ago

Why is this a thing, who uses this? npmjs.com/package/is-even

0 0 2 122 0

class @classpath

a year ago

Are you really a CTA if you dont trade mongolian horse cheese non-deliverable forwards?

0 0 3 159 1

class @classpath

a year ago

When you calculate a trend feature with increasing look back and avg them. ie bolmomAvg = (bolmom8d + .. + bolmom128d)/n Is this not just a rough "exponentially weighted" bolmom? The previous 8 days of data are considered n times, the next 8 n-1 times, etc. @ScottPh77711570

6 1 16 5K 22

class @classpath

a year ago

x: Change in volume y: Return Its generally increasing, but a fat negative bar at the top decile, looks like this decile is catching all the pump and dumps lol. This is crypto perp's btw

0 0 4 226 0

Download Image

class @classpath

2 years ago

Modern ML algos (NN's, Random forest) dont act like simpler algo's in their bias-variance curve. You've seen the U shaped one. But recent literature has shown the "double decent phenomenon" en.wikipedia.org/wiki/Double_de…