21–26 of 26 posts
> However, these calculations did not take into account the effects of “radiative corrections” – the continuous emission and re-absorption of short-lived “virtual particles” (see box) by the electron or muon – which increases g by about 0.1%. Our universe is a subtle mess.
The data-driven approach yields an accurate SM prediction, but the universe and lattice methods agree on a "bad" SM result, due to an undiscovered fact about measure theory and the correct interpretation of path integrals that the implementations of lattice QCD accidentally get right. Since the limit taken by lattice methods have never been proven to equal what we think the right interpretation of path integrals should be, the former could be right when the latter is erroneous.
Weren't you guys supposed to save these stories for Muonday Mondays? Weekend's too long of a wait?
I miss Topological Tuesdays or Turing Thursdays
https://azure.microsoft.com/en-us/blog/quantum/2025/02/19/mi...
Since nobody with expertise is speculating about the reason, here is some speculation from somebody without expertise: The data-driven approach yields an accurate SM prediction, but the universe and lattice methods agree on a "bad" SM result, due to an undiscovered fact about measure theory and the correct interpretation of path integrals that the implementations of lattice QCD accidentally get right. Since the limit…
Lattice QCD methods are definitely proven to be equivalent to the continuum theory when you take the appropriate limits. This isn't some mathematical mystery - Wilson's groundbreaking work in the 70s established this, and it's been rigorously developed since.
The actual g-2 puzzle is much more straightforward: we have competing measurements from different sources:
1. The Fermilab Muon g-2 experiment in Illinois directly measured the muon's anomalous magnetic moment.
2. The "data-driven" prediction used e+e- collision measurements from various experiments (BaBar at SLAC, KLOE in Italy, etc.) to calculate what the Standard Model predicts for g-2. This initially disagreed with Fermilab's measurement.
3. Lattice QCD calculations (like those from the BMW collaboration) produced theoretical predictions that matched Fermilab's experimental result.
4. More recently, new e+e- measurements from the CMD-3 experiment at VEPP-2000 in Russia (2023) are bringing the data-driven approach closer to both the lattice calculations and Fermilab's experimental values.
So it's really just a question of which experimental measurements of e+e- collisions are most accurate/resolve any experimental differences, not some exotic mathematical issue with QFT formulations.
Since nobody with expertise is speculating about the reason, here is some speculation from somebody without expertise: The data-driven approach yields an accurate SM prediction, but the universe and lattice methods agree on a "bad" SM result, due to an undiscovered fact about measure theory and the correct interpretation of path integrals that the implementations of lattice QCD accidentally get right. Since the limit…
Former physicist here. Your speculation about path integrals and measure theory isn't quite right. Lattice QCD methods are definitely proven to be equivalent to the continuum theory when you take the appropriate limits. This isn't some mathematical mystery - Wilson's groundbreaking work in the 70s established this, and it's been rigorously developed since. The actual g-2 puzzle is much more straightforward: we have c…
> However, these calculations did not take into account the effects of “radiative corrections” – the continuous emission and re-absorption of short-lived “virtual particles” (see box) by the electron or muon – which increases g by about 0.1%. Our universe is a subtle mess.
yeah... who designed this thing?