Decision Hygiene: A Good Decision Is Not a Good Outcome

Property hands you one observation per decade. With a sample of one you cannot learn from outcomes: a sloppy process that met a rising market is indistinguishable from skill. So grade the process, and build the apparatus before you need it.

The pre-mortem. Before signing, both of you write separately, 20 minutes, to the same prompt: "It is August 2031. This purchase has been a clear mistake. Tell the story of how." Klein's point is that prospective hindsight – assuming the failure has already happened – surfaces far more concrete causes than "what could go wrong?" Expect: a €14,000 façade Sonderumlage; a tenant who stopped paying in month 8; Anschlussfinanzierung at 6%; one of you leaving tech; €12,000 of work when the flat emptied. Then convert each into a number and put it in the stress case from the sensitivity article.

The decision journal. One page per property, written before the offer:

• The three facts that would make this a good decision
• The three that would make it a bad one
• Our maximum price, and where that number came from
• Our forecasts: 2036 price, total months vacant over ten years, refinancing rate
• Confidence, in percent

Date it. Do not revise it. Re-read it at every Eigentümerversammlung. It is the only real defence against hindsight bias, and it is also how two engineers avoid quietly rewriting joint history.

The walk-away checklist – written before viewing, agreed by both, no clause renegotiated in the moment:

1. Price above our pre-computed maximum → walk.
2. Rücklage under €25/m² with no funded plan for roof, heating or façade → walk or reprice.
3. Missing Protokolle or Beschlusssammlung for three years → walk.
4. Combined after-tax burn above €400/month at +1pp → walk.
5. Cash reserve after completion below €15,000 → walk.
6. Any request to sign within 48 hours → walk.

Good decision ≠ good outcome. If prices rise 5%/year this flat will look brilliant and you will have learned nothing about your judgement. If it falls 10%, a sound process was still sound. The only question you can honestly answer in 2036 is: given what we knew and could have known in 2026, and the risk we could actually bear, was this a defensible allocation? Write your answer now, while honesty is still cheap.