FORECASTING = BETTER GUESSES
Not vibes. Not manifesting. Just slightly smarter predictions.
You’re already guessing about the future every single day. Forecasting gives you a system so you guess less wrong.
- Will this job still exist in 5 years?
- Is this trend hype or actually real?
- Should I save, spend, or wait?
THEY DON’T PREDICT.
THEY UPDATE.
Your brain loves shortcuts. Shortcuts feel like confidence. But confidence ≠ accuracy.
Think in percentages, not certainties. “Probably” beats “definitely.”
Break big questions into smaller, answerable ones.
Change your mind when new info arrives. That’s strength, not weakness.
Hot takes spread faster than truth. Forecasters pause. Everyone else posts.
Instead of: “This will definitely work.”
“There’s a 60% chance — and here’s what could change that.”
That shift alone makes you smarter than most rooms.
Replace Vague Words With Numbers
Words like “probably,” “likely,” and “might” are vague-verbiage — they mean different things to different people. Assigning a number forces precision. 60% means 60%, not “kinda yes.”
Use the Full Range
Most things are neither 0% nor 100%. Very few events are impossible or guaranteed. Using the full scale — 5%, 35%, 72% — forces you to commit to a specific belief, which you can then test and update.
Match Confidence to Reality
Calibration means your 70% predictions should actually happen about 70% of the time — not 40%, not 100%. Calibrated forecasters are more accurate than confident ones, even when less certain.
State the question precisely
Vague questions produce vague estimates. “Will this go well?” is not a question. “Will I get a callback within 7 days?” is.
Find the base rate
How often do things like this happen? If 40% of job applications get a callback, start at 40%. This is your anchor — the “outside view.”
Adjust for specifics
Now consider what makes your situation different. Stronger resume? Move up to 55%. No network connection? Move down to 30%. Small, reasoned adjustments only.
Commit to a number
Write it down. Saying “I think there’s a 55% chance” out loud makes you accountable to your own reasoning — and gives you something to update later.
Track your outcomes
Over time, compare what you predicted to what happened. This feedback loop is how calibration improves. You can’t tune an instrument you never listen to.
Biases That Break Probability Thinking
Your brain fights against accurate probability estimates. These are the main attackers:
Assigning 90%+ to things that only happen 60% of the time. The loudest internal voice inflates estimates.
ANCHORINGThe first number you hear sticks — even if it’s irrelevant. Always check: am I adjusting enough from my anchor?
AVAILABILITY BIASIf you can easily remember a dramatic example, you overestimate how often it happens. Dramatic ≠ common.
BINARY COLLAPSEDefaulting to “it’ll happen” or “it won’t.” The space between 5% and 95% is where most of reality lives.
Question: “Should I move to a new city for this job offer?”
Vague thinking: “It’ll probably work out.” (Could mean 55% or 90% — no way to check.)
Probability thinking: “There’s a 65% chance the role is as described, a 50% chance I build a strong network within a year, and a 30% chance I’d want to move back within 18 months.” Now you can actually reason about it — and update each estimate as you learn more.
Fermi-Style Thinking
Named after physicist Enrico Fermi, who estimated the number of piano tuners in Chicago from scratch using only known quantities and reasoned steps. You expose what you don’t know — and that’s the point. Ignorance made visible is ignorance you can address.
Precision Through Parts
A big question carries all its uncertainty in a single lump. When you split it into components, the errors in each part are smaller and often cancel each other out. Combining rough estimates from five parts is more accurate than one rough guess at the whole.
Separate Knowable From Unknown
Every question has parts you can find data on and parts that remain genuinely uncertain. Sort them into two buckets. Work the knowable parts first — they shrink the unknown parts, and sometimes eliminate them entirely.
State the question in one sentence
“Will AI replace my role within five years?” Write it out exactly — precision here prevents confusion later.
List every sub-factor you can think of
What drives the answer? For the AI question: task automation rate in your field, pace of adoption in your industry, reskilling speed, economic incentives for companies.
Sort: knowable vs. genuinely uncertain
Automation rate in your field — knowable (there’s research). Your company’s specific adoption timeline — uncertain. Work the knowable pieces first.
Estimate each component separately
Assign a number or range to each knowable sub-factor. Even rough estimates per component produce a more defensible overall answer than one gut shot.
Combine and check
Multiply or weight your estimates into a final probability. Then ask: does this pass a basic sanity check? If not, find where your sub-estimate is off.
Big question: “Should I go freelance full-time?”
Decomposed into: (1) Do I have 3+ months of savings runway? (2) Do I have at least one client lined up? (3) Is my field one where freelance work is available? (4) What’s my monthly burn rate vs. average freelance income in my field?
Each of these is answerable with real data. Answering all four gives you a far better read than asking the big question all at once.
Not All New Info Qualifies
An update is warranted when new information is genuinely new, relevant, and independently sourced. Seeing the same opinion five times in your feed is not five data points — it’s one, repeated. Check: does this info change the actual probability, or just my feelings about it?
Proportional, Not Dramatic
Small signals = small moves. A single news headline might shift 65% to 62%. A major official announcement might shift 65% to 45%. The size of your update should match the weight of the evidence — not your emotional reaction to it.
Stubbornness vs. Flip-Flopping
Two ways to update badly: never update (ego defense — treating old beliefs as identity) or update on everything (overreacting to noise, chasing the latest story). Both kill accuracy. The goal is steady, reasoned recalibration.
State your current estimate clearly
Before consuming new information, write down your existing probability. “I think there’s a 70% chance this project gets funded.” This creates your baseline.
Evaluate the new information independently
Ask: Is this genuinely new? Is it from a reliable, independent source? Is it directly relevant to my question — or just adjacent to it?
Ask what it implies about your estimate
If this information is true, does it make my event more likely, less likely, or equally likely? By roughly how much? Name the direction and magnitude before moving the number.
Adjust proportionally
Move the number. If the evidence is moderate, move modestly (±5–10%). If it’s substantial and direct, move more (±15–25%). Resist the pull to swing to extremes.
Log your update and its reason
A short note — “moved from 70% to 58% because the lead investor passed” — builds your reasoning record. This is how you learn to update better over time.
The Signal/Noise Filter
New, independent data contradicting your assumption
A direct outcome that resolves one of your sub-questions
Expert consensus shifting significantly in one direction
A structural change to the situation (new law, key person leaving)
Social media volume or sentiment around the topic
Opinions from people with no more information than you
Information that just confirms what you already believed
Your emotional reaction to a headline or outcome
Estimate: “70% chance I get a job offer within 60 days.”
Valid update trigger: You get an interview rejection with specific feedback that your salary expectations are too high for the market. → Move to 52%. Reason: structural barrier identified.
Invalid update trigger: You see a viral post saying “the job market is terrible right now.” → No update. That’s social signal, not data specific to your situation. Resist the noise.
System 1 vs. System 2
System 1 runs constantly, decides instantly, and operates on pattern recognition and emotion. It’s efficient — and regularly wrong on complex problems. System 2 is deliberate and effortful. It kicks in when you recognize that a situation requires actual analysis, not just reflex.
Algorithms Reward Speed
Social media punishes the pause. Engagement metrics reward the fastest, most confident take. But speed and accuracy are in direct conflict under uncertainty. The forecaster’s edge is choosing accuracy over the appearance of confidence — which requires resisting the pull to react instantly.
Not Paralysis. Not Waffling.
Pausing is not indefinite delay or chronic indecision. It’s a brief, deliberate intervention — seconds to minutes — that lets you notice whether your first reaction is System 1 reflex or reasoned judgment. Then you act. Faster decisions on low-stakes questions. Slower on high-stakes ones.
Notice the reflex forming
Before you share, reply, or decide — observe your first reaction. Is it strong and immediate? That’s System 1 activating. Label it: “That’s my gut. Not yet my answer.”
Ask: is the stakes level high?
Low stakes (what to eat, quick reply to a friend): let System 1 run. High stakes (career move, public statement, financial decision): invoke the pause deliberately.
Introduce a structural delay
Draft the message, then wait 10 minutes before sending. Write the decision down before executing it. Sleep on it if it involves money or relationships. Physical time is the simplest pause tool.
Ask one adversarial question
“What would have to be true for me to be wrong right now?” This single question activates System 2 and disrupts confirmation bias — your brain’s tendency to only look for evidence that agrees with you.
Then decide — and track
The pause is not the destination. Make your call, record your reasoning briefly, and move. Revisit later to see if the pause improved your outcome. It almost always does.
What Happens Without It
CONFIRMATION BIAS
You search for evidence that confirms your first reaction — and find it. The pause forces you to also look for evidence that challenges it.
AFFECT HEURISTIC
Strong emotion (fear, excitement, disgust) hijacks probability estimates. You overweight bad outcomes when scared, good outcomes when excited. The pause lets the emotion pass before the judgment forms.
NARRATIVE FALLACY
Your brain prefers a clean story over a probabilistic one. The pause creates enough friction to ask: “Is this a story, or is this actually what’s happening?”
RECENCY BIAS
Whatever happened most recently feels most important. The pause activates the question: is this recent information actually more relevant, or just more vivid?
Scenario: You see a news story saying a company you invested in just had a bad earnings report.
System 1 (no pause): Immediate panic. Sell immediately. Share the story with commentary: “I knew it.”
System 2 (after the pause): One quarter of bad earnings — what’s the base rate for recovery in this sector? Is this structural or temporary? Does this change my 18-month thesis, or just one data point? What would need to be true for this to be a buying opportunity instead?
The pause doesn’t guarantee the right call. It guarantees the reasoning was yours — not your reflex’s.
CONFIDENCE IS NOT ACCURACY
The loudest voice is rarely the best guess. Algorithms reward certainty. Reality rewards flexibility.
- Locks in early
- Defends the ego
- Ignores new data
- Performs certainty
- Stays curious
- Stays humble
- Stays adjustable
- Thinks in probability
1-MINUTE FORECASTING HABIT
Before any decision. No spreadsheets. No genius required. Just better life odds.
Leave a Reply