What if the table says one thing but I still feel the other option is right?

Take the feeling as a new criterion and score it explicitly. If you cannot score it, the feeling is either sunk cost or default bias — run the zero-history test.

Do I really need to write all of this down for a small decision?

No — reversible calls under 30 days can be decided in one sitting. Write the record only when the decision is expensive to reverse or when you know you will want to revisit it.

How many criteria is too many?

Six is the practical ceiling. Past eight rows the weights stop carrying meaning and the total becomes noise — cut the criterion that changes the winner least.

I already started down one path and it is going badly — should I switch?

Run the zero-history test: ignoring everything already spent, which option would you choose today? Sunk cost is the most common reason people stay on a failing path.

Compare options, decide once

Ohmaybe — How to make decisions you will not regret | ohmaybe.com

Most regretted decisions were not decided badly — they were framed badly. This reference walks through the full loop: define the real question, list constraints, weigh options on explicit criteria, and record enough to revisit the call later.

A map of this guide's sections

Frame before you compare

Name the real question, write the goal as a measurable sentence, and separate hard constraints from preferences. Reversibility sets the time budget for everything that follows.

Compare with weights, not vibes

A six-row criteria table with weights set before scoring turns an argument into arithmetic. Flip one weight by two points to check that the winner is robust.

Check traps, then record

Sunk cost, default bias, and analysis paralysis each have a cue and a countermeasure. Write a one-page record with a revisit condition so the decision can be revisited rather than re-argued.

What is the real question behind your choice?

Before comparing anything, name the goal, the constraints, and whether the decision can be reversed.

Most stalled decisions are stuck because the question on the table is not the question that matters. "Should I take this job?" hides "Do I need income now, or do I need a specific skill in 18 months?" — and the two questions have different answers.

Write the goal as one sentence with a measurable end state: "Reduce my commute from 90 minutes to under 25 minutes per day." A goal without a number cannot be compared against options, because there is nothing to check the option against.

Then list constraints as hard filters, not preferences. A hard constraint eliminates an option outright; a preference only shifts its score. Mixing the two is the most common reason a comparison table comes out feeling wrong even when the arithmetic is right.

Goal: one sentence, one measurable end state.
Constraints: what eliminates an option immediately, written before you look at the options.
Reversibility: can you undo this in 30 days for less than it cost to start? If yes, decide faster.

Checklists by stage

Framing

Goal sentence

One sentence, one measurable end state. If the goal has no number, the comparison has nothing to check the options against.

Framing

Constraint filter

Write hard constraints before listing options. A constraint eliminates; a preference only shifts a score.

Comparing

Weight once

Set criterion weights before scoring. Adjusting weights after seeing totals is moving the goalposts.

Comparing

Robustness flip

Change one weight by ±2 points. If the winner flips, the options are close — pick the reversible one.

Traps

Zero-history test

Ask: starting fresh today with no prior investment, which option would I choose? Sunk cost fails this test.

Recording

Revisit trigger

A measurable condition — "commute over 40 minutes for two weeks" — not a mood. Written the same day you decide.

If you cannot state the reversal condition, you are not deciding — you are hoping.

Changing weights after seeing scores is the comparison equivalent of moving the goalposts.

Real objections, answered

What if the table says one thing but I still feel the other option is right?: Take the feeling as a new criterion and score it explicitly. If you cannot score it, the feeling is either sunk cost or default bias — run the zero-history test.
Do I really need to write all of this down for a small decision?: No — reversible calls under 30 days can be decided in one sitting. Write the record only when the decision is expensive to reverse or when you know you will want to revisit it.
How many criteria is too many?: Six is the practical ceiling. Past eight rows the weights stop carrying meaning and the total becomes noise — cut the criterion that changes the winner least.
I already started down one path and it is going badly — should I switch?: Run the zero-history test: ignoring everything already spent, which option would you choose today? Sunk cost is the most common reason people stay on a failing path.

Method and limits

No external sources are cited; the content is a self-contained reference on decision mechanics.

A map of this guide's sections