I honestly didn't do it for trading. I did it because PlanB's STFX model interested me but is like ok you can't do a regression on his stuff and get a statistically significant model for BTC using daily pricing. He did stuff where it was like grouping things together for certain time periods gave you statistical significance but it was fuzzy math.
I watched some thing where people were sliding gm2 money supply around and matching peaks and troths with Bitcoin. Really rudimentary stuff. But I was like huh, it does look like btc lags gm2 but it was obvi not directly corelate.
I started thinking about how scarce assets go up in value first when there is inflation; inflation is a result of M2 creation. I wonder if I can sharpen up S2F and prove to myself it is the legit driver of BTC value.
Then when I got a bump on my first try I started tinkering around with lag because M2 doesnt create inflation overnight. Then finding the best lag. Then plotting r2 value over different lag times. Once I optimized GM2 lag I was like we'll what about lag on S2F. That tightening shouldn't be felt overnight either.
As you can see once I start digging into something I can't stop. Messed with weighting and smoothing and lag and optimized each. I think I could get it even tighter if I used daily pricing instead of weekly. But .97 is very high and it probably would still round to .97 no matter what.
Had to share because I thought it was cool and I thought it was weird that optimized smoothing and lag all were fibs.
jasontheoriginal / Jason the Original
npub1fg…nyj2h
2025-04-11 21:09:13
in reply to nevent1q…uelw
Author Public Key
npub1fgnuaulzmkpnpe8jy843hf7lhwefh5xyejl5s4307ajw25tgrlssxnyj2hPublished at
2025-04-11 21:09:13Event JSON
{
"id": "b49a9d2453436fa9a85febfbafd8e47c06a151831d2b5624bcf57ffae6b01816",
"pubkey": "4a27cef3e2dd8330e4f221eb1ba7dfbbb29bd0c4ccbf48562ff764e551681fe1",
"created_at": 1744405753,
"kind": 1,
"tags": [
[
"e",
"ca447db2a8e50071db6bda4d9045bf3fc3a4d926d37216271b0e2270209b6b5b",
"",
"root"
],
[
"e",
"625f281e644102b41ca3ac5df26a7f03957f5a49a22cacdbe09d43c6ffcdc36a",
"",
"reply"
],
[
"p",
"fcf6fee0e959c7195dadc5f36fe5a873003b389e7033293b06057c821fcbc9c5"
],
[
"p",
"4a27cef3e2dd8330e4f221eb1ba7dfbbb29bd0c4ccbf48562ff764e551681fe1"
],
[
"p",
"61854976ba1a538b9dfed1b1df9b42e57602790e9354c59891be7b3713db637c"
]
],
"content": "I honestly didn't do it for trading. I did it because PlanB's STFX model interested me but is like ok you can't do a regression on his stuff and get a statistically significant model for BTC using daily pricing. He did stuff where it was like grouping things together for certain time periods gave you statistical significance but it was fuzzy math.\n\nI watched some thing where people were sliding gm2 money supply around and matching peaks and troths with Bitcoin. Really rudimentary stuff. But I was like huh, it does look like btc lags gm2 but it was obvi not directly corelate. \n\nI started thinking about how scarce assets go up in value first when there is inflation; inflation is a result of M2 creation. I wonder if I can sharpen up S2F and prove to myself it is the legit driver of BTC value. \n\nThen when I got a bump on my first try I started tinkering around with lag because M2 doesnt create inflation overnight. Then finding the best lag. Then plotting r2 value over different lag times. Once I optimized GM2 lag I was like we'll what about lag on S2F. That tightening shouldn't be felt overnight either.\n\nAs you can see once I start digging into something I can't stop. Messed with weighting and smoothing and lag and optimized each. I think I could get it even tighter if I used daily pricing instead of weekly. But .97 is very high and it probably would still round to .97 no matter what.\n\nHad to share because I thought it was cool and I thought it was weird that optimized smoothing and lag all were fibs.",
"sig": "70e0be107c74af2d0a68327ed42eabb24557bd6ba0b29b172a4c075c003c3aea25f454b6fe388f701079b3c3ccdf5efd317caf1931698a05081fd99b8e77fb99"
}