Hey Stardance crew! Three months since my last devlog, so grab a snack and let me catch you up.

DOSEFIELD estimates how much radiation an astronaut gets behind shielding on a trip to Mars, then checks itself against real data from NASA’s Curiosity cruise (MSL/RAD).

The big news: I found a mistake in my own results. In my first public version, the solar modulation value (W = 30) could not be reproduced from the derivation I had written down. So I went back to the Oulu neutron-monitor data, matched the exact time window RAD used, skipped the solar-storm days, and got W = 47.50.

The new numbers look worse, and I kept them:

• absorbed dose: 0.264 vs 0.458 mGy/day measured (0.58x)
• dose equivalent: 1.318 vs 1.75 mSv/day (0.75x)
• mean quality factor: 4.995 vs 3.82 (1.31x)

My model skips secondary particles and neutrons, so it under-predicts the dose. I report the gap and leave it alone.

Also shipped: a check against the full NIST PSTAR table (max deviation 1.55% from 10 to 1000 MeV), 190 passing tests, a code archive on Zenodo (DOI 10.5281/zenodo.23433384), a revised paper on SSRN, and a website that links to all of it.