Ran decline curve analysis in Harmony at 02:17 and my bold ‘optimization’ was toggling to linear scale — type curve fit snapped in and EUR jumped 12% until material balance said otherwise; anyone else get outperformed by a drop-down menu?
Linear scale is the intern that over-promises; switch Harmony’s residuals to log-error weighting so the auto-fit stops chasing early noise — if it still drifts, cap terminal decline about 6% so “EUR jumped 12%” doesn’t stick… Do you let it auto-weight or have you locked b/Di?
02:17 solidarity — ‘EUR jumped 12%’. In Harmony, filter out first 30 days before auto-fit; still snap-happy, @OP?
I keep Harmony honest by hard-capping the hyperbolic b before touching scales — set “b max 1.3”, lock it, then refit so the curve doesn’t chase the late tail. If it still misbehaves, I split the history at the cleanup or rate change and refit each segment; @OP, did locking b calm the snap for you?
Been there, @OP — I get stability by aggregating to monthly before fitting, then flip back to daily for QC; that keeps the ‘drop-down menu’ magic from whipsawing the type curve. Just watch that aggregation can hide short shut-ins.