Skip to content after assessment

Assessment loading

If the interactive assessment does not start, you can still read the methodology and evidence below, then try again later.

Browse all assessments

Where Are Your Calibration Weak Points?

Your Probabilistic Thinking Profile

Where Are Your Calibration Weak Points?

Item answers and results stay on this device

A structured self-reflection, not a diagnostic.

A 20-scenario self-report that groups your selected answer points across five editorial dimensions. It is not a validated measure of an underlying trait.

20 scenarios5 dimensions4 forced choicesno neutral option

Topic context from the cited source; it does not validate this questionnaire or its score thresholds.

Tversky & Kahneman (1974)

What your answers coverConfidence CalibrationExpressing your uncertainty as honest odds or ranges, then checking against outcomes.Base-Rate SensitivityWeighing statistical averages over a single vivid example when estimating likelihood.Inside vs Outside ViewAnchoring your forecasts on how similar past efforts turned out, not just your plan.Bayesian UpdatingShifting your beliefs in proportion to how strong new evidence actually is.Overconfidence PatternConsistently overestimating your ability, speed, or accuracy compared with how you actually perform.

Calibration Test: How Well-Calibrated Are Your Predictions?

Twenty forced-choice scenarios are summed into editorial answer-score bands; the result does not diagnose or measure an underlying trait.

~7-9 min · Item answers and results stay on this device · Links created by this assessment contain no scores

Answer with your real predictions in mind — the ones you got wrong count double here, and that's the point.

This assessment is a structured self-reflection tool and educational resource — it is not a validated diagnostic and is not a substitute for clinical evaluation. Every dimension insight references peer-reviewed literature for educational context only. If a result here raises real concerns for you, take it to a qualified professional — a self-assessment is a prompt for reflection, not a verdict. Self-report instruments of this kind have no validated retest interval; the four-to-six-week Second Reading is a sensible practice, not a clinical standard.

Welcome back — you have a profile in progress.

Your profile

0 / —

Profile

Dimension profile

The Reading

Strongest dimension

Learn more →

Open your assessment summary

Dimension log

This assessment is a structured self-reflection tool and educational resource — it is not a validated diagnostic and is not a substitute for clinical evaluation. Every dimension insight references peer-reviewed literature for educational context only. If a result here raises real concerns for you, take it to a qualified professional — a self-assessment is a prompt for reflection, not a verdict.

Your answers

Calibration Test: How Well-Calibrated Are Your Predictions?

This literature-informed structured self-reflection adds up what you select about five dimensions of probabilistic thinking — how closely your confidence tracks actual outcomes, whether you lean on base rates or vivid cases, how consistently you adopt the outside view, how you update beliefs from new evidence, and which forms of overconfidence show up in your answers. It does not measure your actual prediction accuracy.

Interactive scoring requires JavaScript. You can still read every question and answer option below.

When you express high confidence in a fact or forecast — saying you are very sure something is true — how often do you track whether you were actually right?
- Almost never — I rarely revisit whether I was right
- Only when an outcome is particularly striking or obvious
- Sometimes — I keep rough track in certain domains
- Actively — I log predictions and check outcomes
When you do not know something for certain, how do you typically communicate that uncertainty to others?
- I typically give a definite answer even when I am not sure
- I say 'probably' or 'likely' without quantifying
- I signal uncertainty verbally but rarely with a number or range
- I give a range or rough probability — '70% likely' or 'between X and Y'
When you give someone an estimate for a cost, timeline, or quantity, what does your answer usually sound like?
- A single confident number — naming a range instead would make me sound like I don't trust my own judgment
- I work out a range in my head, but I still report just the midpoint — a clean number lands better than a wishy-washy range
- I'll offer a range if pressed, though my default reported range still tends to run narrower than it probably should
- I state my honest range up front, even though a single tidy number would sound more decisive
After a major decision or forecast that has now resolved, how often do you deliberately revisit what you predicted versus what actually happened?
- Almost never — once something is over, I move on
- Occasionally, when an outcome surprises me enough to prompt reflection
- Sometimes, informally — I note when I was badly off but keep no records
- Regularly and deliberately, reviewing written records of predictions against outcomes
When a striking individual story and the relevant statistics point in different directions, what does your honest reaction sound like?
- The real story wins — numbers can't capture what an actual case tells you, and I trust what's in front of me over an abstract statistic
- I lean toward the story, but I'll mention the statistic so I don't sound like I'm ignoring the data
- I try to hold both in mind, though the vivid case still pulls harder than the numbers probably justify
- I let the statistic lead and treat the story as one more data point, even when that feels colder than trusting my gut
Before committing to a project or forecast, how often do you seek data on how often similar efforts have succeeded or failed?
- Rarely — every situation feels unique enough to make past data unhelpful
- Occasionally, if historical data is easy to find
- Often, but I treat prior data as background rather than the main anchor
- Consistently — I find base rates for similar situations before forming my estimate
After hearing a vivid account of something going dramatically wrong (or unexpectedly right), how much does it shift your estimate of how common that type of event is?
- It shifts me substantially — that case now feels highly representative
- It shifts me noticeably even though I know one case is not strong evidence
- It shifts me a little, but I make a note to look for overall rates
- It barely shifts me — I immediately ask how common this is in the full population
When you encounter a probability expressed as a small or large number (e.g., a one-in-a-hundred chance), how readily do you translate it into a concrete frequency format to check whether your intuition matches the number?
- Rarely — I react to the number without converting it
- Sometimes, if the probability seems surprising
- Often, especially for small or large probabilities
- Habitually — I reframe percentages as concrete frequencies to anchor my intuition
When you're forming a forecast for a project, goal, or outcome, what does your actual process look like?
- I build the plan from what I know about this specific situation — leaning on other people's outcomes would just water down my own judgment
- I'll glance at how similar efforts went, but I mostly trust my own read of the specifics
- I do check comparable efforts, but usually after I've already landed on my own estimate
- I look up how similar efforts actually turned out and anchor there first, even though it can feel like I'm second-guessing my own read of the situation
When you discover that similar past efforts in your domain have a high failure rate, how much does that actually change your confidence in your own current effort?
- Barely — I believe mine is different enough that the historical rate does not apply
- A little — it gives me mild concern but I still rely mainly on my own assessment
- Noticeably — I lower my confidence meaningfully, though not fully in line with the rate
- Substantially — I treat the failure rate as strong prior information and anchor on it
When estimating how long a complex effort will take, how explicitly do you anchor on how long similar past efforts have actually taken — rather than reasoning from your plan's specific steps?
- I reason almost entirely from the specific steps in my plan
- I use my plan as the primary anchor and make minor adjustments for what I have heard about similar efforts
- I try to weight both, though step-based reasoning tends to dominate
- I start with the reference class average, then adjust for the specifics of my case
When your own forecast for a project is more optimistic than the historical average for similar projects, how do you respond?
- I trust my own assessment — I know more about the specifics than the average captures
- I take mild note but do not revise much — I believe I am controlling for the factors that caused past failures
- I revise downward somewhat and try to identify specific factors that make my case genuinely different
- I treat the gap as a strong warning signal and substantially raise my estimate of risk or time required
When solid evidence comes in that contradicts a belief you hold, what does your honest reaction look like?
- I hold my ground — constantly revising your position over every new piece of evidence just makes you look like you never had convictions in the first place
- I'll shift a little to acknowledge the evidence, but I try not to move far off a position I've thought through
- I do update meaningfully, though I notice I sometimes hold on longer than the evidence really justifies
- I shift my position in proportion to the evidence, even when that means reversing a view I stated confidently — even if it looks like I can't make up my mind
When new information is mixed — partly supporting and partly challenging your current view — how do you typically update your beliefs?
- I focus on the parts that support my view and discount the challenging parts
- I notice both sides but tend to settle on my existing view with only minor revision
- I try to weight the mixed signals but find it hard to avoid anchoring on my original view
- I extract what each piece of evidence independently implies and then integrate both signals
When you receive a new piece of information, how explicitly do you consider the reliability of its source before deciding how much to update your views?
- I mainly react to the content — source reliability rarely enters my thinking explicitly
- I notice source reliability sometimes but it rarely changes how much I update
- I often adjust for source reliability, but not always in a principled way
- I explicitly assess source reliability and scale my update accordingly before revising any belief
After updating a belief based on new evidence, how often do you actively seek out evidence that might push back in the other direction?
- Almost never — once I have updated, I move on
- Only if the update was large or particularly surprising
- Sometimes, though more often when updating toward a less comfortable conclusion
- Routinely — I treat each significant update as a prompt to stress-test the updated belief
On skills or tasks you consider yourself above average at, how accurately do you know your actual ranking compared to others?
- I rarely check — I assume I am in the top tier until proven otherwise
- I sometimes compare with specific others but tend to benchmark against my weaker peers
- I try to assess fairly but notice I tend toward optimism in my relative estimates
- I deliberately look for performance data and compare against a broad, representative sample
When you estimate how long a task will take, what does your honest estimate actually sound like?
- I give myself an aggressive, tight timeline — setting a padded estimate just gives you permission to be slow
- I know my estimates run tight, but I keep giving ambitious numbers because I think it's what drives the pace
- I pad my estimates a little after noticing they usually run short, though I still lean toward the ambitious end
- I estimate based on how long similar tasks actually took before, even when that number feels unambitious compared to what I'd like to promise
When you give a range you are 'fairly confident' contains the true answer, how often does the true answer actually fall outside your stated range?
- Very often — the true answer is outside my range much more than I expect
- Often — my ranges are too narrow, though not dramatically so
- Occasionally — I am slightly overconfident in my ranges
- Rarely — my stated ranges capture outcomes at close to the rate I claim
On genuinely difficult questions or tasks — ones you find challenging — how does your stated confidence level compare to how accurate or successful you actually end up being?
- I am far more confident than my accuracy warrants — difficulty does not dampen my confidence much
- I am somewhat more confident than the results suggest, but I do notice when tasks are hard
- My confidence and accuracy track reasonably well, though I lean slightly overconfident on hard tasks
- On hard tasks my confidence drops proportionally — I know when I genuinely do not know

Appendix · Method & evidence Decisions · Case file

How this instrument works

This is a 20-item structured self-assessment of your calibration weak points, read across 5 dimensions: Confidence Calibration; Base-Rate Sensitivity; Inside vs Outside View; Bayesian Updating; and Overconfidence Pattern. Each item is a concrete scenario with four forced-choice responses — no neutral midpoint, so every answer commits to a direction — and response order is shuffled on every run to vary position; this implementation has not been tested for bias reduction. 5 of the 20 items are reverse-keyed — the attractive-sounding response is the structurally weak one — with wording intended to challenge socially desirable framing; this design has not been tested for bias reduction.

Scoring combines an overall reading with a per-dimension profile. Cascade flags compare configured dimension totals with editorial thresholds. They do not establish a compound risk and may not reflect the item-level pattern. The result bands are editorial guidance, not clinical thresholds: they organise the reading, they do not describe.

Every dimension insight cites peer-reviewed literature, and most answer reveals do — 8 sources back this instrument, each with DOI metadata supplied at build time; source relevance and claim interpretation remain editorial judgments. Compiled by the High Performance Culture editorial team. Instrument v1.1 · August 2026.

Read the full assessment methodology →

Items: 20
Dimensions: 5
Format: Forced choice, 4 options
Sources: 8
Instrument: v1.1 · August 2026

The 5 editorial answer groups

Confidence Calibration

Strength here looks like actively recording predictions and reviewing outcomes to close the feedback loop, so stated confidence levels are anchored in a real track record rather than recollection.

The vulnerable pattern: expressing uncertainty vaguely and rarely revisiting predictions, leaving confidence levels unchecked against actual accuracy.

Sample itemWhen you express high confidence in a fact or forecast — saying you are very sure something is true — how often do you track whether you were actually right?

First structural move: Start a prediction log this week: record your stated confidence and the outcome for three decisions. A month of entries tells you more about your calibration than years of untracked experience.

Go deeperProtocol: The Feedback Loop Protocol →Glossary: Hindsight bias →

Base-Rate Sensitivity

Strength here looks like anchoring probability estimates on statistical base rates and deliberately down-weighting vivid individual cases and memorable events relative to what the data actually supports.

The vulnerable pattern: allowing striking individual examples and easily recalled events to dominate frequency estimates, while statistically valid base rates recede into background.

Sample itemWhen a striking individual story and the relevant statistics point in different directions, what does your honest reaction sound like?

First structural move: The next time a vivid story shifts your sense of how common something is, pause and ask: 'What is the actual base rate for this across the full population?' Look it up before updating your estimate.

Go deeperDeep dive: The Availability Heuristic →Glossary: Survivorship bias →

Inside vs Outside View

Strength here looks like routinely anchoring forecasts on reference class distributions — how often similar efforts actually succeeded — before adjusting for case-specific detail.

The vulnerable pattern: building plans from case-specific steps while treating historical failure rates as inapplicable, sustaining inside-view optimism that the evidence does not support.

Sample itemWhen you're forming a forecast for a project, goal, or outcome, what does your actual process look like?

First structural move: Before your next significant forecast, spend 20 minutes in the outside view: find 10–20 comparable past efforts and record their outcomes. Use that distribution as your starting point, then name the specific differences that justify any upward adjustment.

Go deeperProtocol: The Mental Models Protocol →Glossary: Optimism bias →

Bayesian Updating

Strength here looks like revising beliefs in proportion to evidence strength — treating disconfirming and confirming evidence with similar scrutiny, and explicitly accounting for source reliability.

The vulnerable pattern: underweighting disconfirming evidence and relying on selective processing to preserve existing beliefs, with source reliability inconsistently factored in.

Sample itemWhen solid evidence comes in that contradicts a belief you hold, what does your honest reaction look like?

First structural move: The next time you encounter solid evidence contradicting a belief you hold, write down: the belief, the evidence, and what an uninvested observer would conclude from that evidence alone. Then update by that amount.

Go deeperProtocol: The Probabilistic Thinking Protocol →Glossary: Bayes' theorem →

Overconfidence Pattern

Strength here looks like calibrating confidence to difficulty — letting stated certainty drop proportionally on hard or uncertain tasks, and checking relative ranking against representative data rather than intuition.

The vulnerable pattern: maintaining high confidence on difficult tasks, underestimating time and effort, and setting intervals that are too narrow — showing overprecision, overestimation, or overplacement.

Sample itemOn skills or tasks you consider yourself above average at, how accurately do you know your actual ranking compared to others?

First structural move: Choose one task you feel confident about and find objective performance data. Compare your actual result to the distribution of similar performers. Use that gap as a calibration anchor for your confidence in related domains.

Go deeperDeep dive: The Overconfidence Effect →Glossary: Overconfidence bias →Glossary: Dunning-Kruger effect →

Questions, answered

Is this a diagnosis?

No. This assessment is a structured self-reflection tool, not a validated clinical instrument. Its result bands are editorial guidance, not clinical thresholds — they organise your answers into a readable profile; they do not describe a condition. If you have concerns about your health or wellbeing, speak to a qualified professional.

How were the questions designed?

Each of the 20 items presents a concrete scenario with four graded responses and no neutral midpoint, so every answer commits to a lean. Items were authored against the peer-reviewed literature listed in the evidence file on this page, and citation metadata is supplied at build time; source relevance and claim interpretation remain editorial judgments. Response order is shuffled on every run to vary position; this implementation has not been tested for bias reduction.

What happens to my answers?

Assessment answers stay on this device. Unfinished progress expires after 30 days. A completed profile and optional comparison history remain until you delete local assessment data or clear browser data. If you opt into email, your email address is sent; item answers are not. If you share a result link, total and dimension scores are included in the URL and may appear in browser, server, CDN, or analytics logs.

Can I share my result?

Links created by this assessment contain no scores. Older result URLs may contain total and dimension scores and can appear in browser history and server or CDN logs; avoid sharing them.

How often should I retake it?

The instrument keeps your previous reading and sets the two side by side when you retake — the Second Reading. There is no validated retest interval; repeat only if a raw answer-point comparison would be useful, and do not treat it as confirmed underlying change.

Why bands instead of percentiles?

Because this assessment does not collect a population score set, there is no reader population to rank you against — and we will not invent one. The 5-dimension profile groups selected answer points for editorial display; it is not a validated structural or decision-utility model.

Where does the evidence come from?

Every dimension insight cites peer-reviewed research, and most answer reveals do — 8 sources stand behind this instrument, each with DOI metadata supplied at build time; source relevance and claim interpretation remain editorial judgments. A few reveals are straightforward editorial reasoning from the same evidence base and carry no citation by design. The full list, with DOIs, is in the evidence file below.

The evidence file

8 peer-reviewed sources stand behind this instrument's items and insights.

Tversky & Kahneman (1974) — Judgment under Uncertainty: Heuristics and Biases. Science. DOI: 10.1126/science.185.4157.1124 ✓ Crossref‑verified
Moore & Healy (2008) — The trouble with overconfidence. Psychological Review. DOI: 10.1037/0033-295x.115.2.502 ✓ Crossref‑verified
Lichtenstein & Fischhoff (1980) — Training for calibration. Organizational Behavior and Human Performance. DOI: 10.1016/0030-5073(80)90052-5 ✓ Crossref‑verified
Kelly & Mandel (2024) — The effect of calibration training on the calibration of intelligence analysts' judgments. Applied Cognitive Psychology. DOI: 10.1002/acp.4236 ✓ Crossref‑verified
Gigerenzer & Hoffrage (1995) — How to improve Bayesian reasoning without instruction: Frequency formats. Psychological Review. DOI: 10.1037/0033-295x.102.4.684 ✓ Crossref‑verified
Flyvbjerg (2007) — Curbing Optimism Bias and Strategic Misrepresentation in Planning: Reference Class Forecasting in Practice. European Planning Studies. DOI: 10.1080/09654310701747936 ✓ Crossref‑verified
Hogarth & Einhorn (1992) — Order effects in belief updating: The belief-adjustment model. Cognitive Psychology. DOI: 10.1016/0010-0285(92)90002-j ✓ Crossref‑verified
Johnson & Fowler (2011) — The evolution of overconfidence. Nature. DOI: 10.1038/nature10384 ✓ Crossref‑verified