Anchoring Bias

A simple explanation

Someone gives you a number — a price, an estimate, a deadline — and your subsequent judgment forms around that number, even when the number was arbitrary, even when you know it was arbitrary. The mind starts at the anchor and adjusts from there, and the adjustment is almost always insufficient. The anchor wins quietly.

This is anchoring bias. Identified by Tversky and Kahneman. The classical experiments showed it with literal random numbers — a spun wheel — pulling judges' estimates of unrelated quantities toward whatever the wheel landed on.

An everyday example

You walk into a furniture shop. The first item you see is a sofa marked at twelve thousand. You feel a small jolt — that is high. Twenty minutes later, you find a sofa you actually want, marked at six thousand. It feels reasonable. You leave with it, slightly pleased at having spent half what the first sofa cost.

The sofa you bought is double what you would have paid had you walked into a different shop with a different first-anchor. The twelve-thousand sofa, which you never seriously considered, did most of the work. The shop's layout — most expensive first — was not accidental. Your reasonable was their anchor.

Why does the first number always win?

Because the mind, asked to produce a numerical estimate, looks for a starting point and adjusts. The first available number is the cheapest starting point, and the adjustment process is bounded — it stops when the verdict feels plausible, not when the evidence would land you. The Threat System, organised around fast verdicts, accepts the bounded process because deeper analysis is more expensive than the gain feels worth.

Knowing about the anchor does not neutralise it. The classical experiments replicate even when subjects are told the anchor is random and instructed to ignore it. The mind starts there anyway. The bias is not about credulity — it is about the architecture of numerical judgment under any kind of uncertainty.

The behavioral loop

The loop runs in seconds:

1. Anchor presented — a number arrives. Price tag, opening bid, asking salary, time estimate.
2. Starting point set — the mind, asked for a verdict, begins at the anchor whether or not the anchor is relevant.
3. Adjustment — the mind moves away from the anchor in the direction the evidence suggests.
4. Premature stop — adjustment stops when the verdict feels plausible, not when the analysis would terminate.
5. Verdict carries the anchor's shape — the final number is closer to the anchor than the evidence alone would have produced.
6. Confidence assigned — the verdict feels considered, because the adjustment process was conscious.
7. No correction — because the anchor was the input the system worked with, there is no obvious moment at which it gets caught.

Emotional drivers

Three quiet drivers:

• The relief of a starting point when the question was open and uncomfortable to hold.
• A felt sense of having done the work, because the adjustment from the anchor was visible to consciousness even though the anchor itself was unexamined.
• A defensive friction when someone proposes a verdict far from the anchor — experienced as the other person being unreasonable, rather than as data about the anchor's grip on your own judgment.

What your nervous system does

Very little autonomically. Anchoring is a cognitive distortion that runs largely below the level of felt signal. The body does not report a spike when an anchor is set; it reports nothing, which is part of what makes the bias hard to catch. The cost shows in the verdicts, not in the moment of judgment.

The Threat System's involvement here is at the level of cognitive economy — the bounded adjustment process is itself a stress-response design, optimised for a world in which deciding fast was more valuable than deciding accurately. In numerical estimation under uncertainty, the design is now mostly distortive.

The DojoWell interpretation

Anchoring bias is a Threat System's cognitive shortcut treating the first available number as the relevant baseline. The substitute is adjust-from-the-anchor; the original ask was estimate-from-the-evidence. They share an outer shape — both produce a number. They share none of the epistemics.

The Meaning Density reading is false_progress. Effort is low per instance and large in aggregate, applied to every numerical judgment under uncertainty. Deposit on accuracy is near-zero when the anchor is irrelevant — the verdict carries the anchor's shape regardless of the evidence. Residue accumulates in the form of overpaid prices, conceded negotiations, distorted time estimates, salaries accepted closer to the offer than the market warranted.

The pattern is especially expensive because the world is full of deliberately set anchors — list prices, opening bids, asking salaries, suggested donations. Each one is an attempt to set the baseline from which your adjustment will be insufficient.

How do I neutralise an anchor?

Three moves:

1. Generate your own number before exposure. What is this worth, what is the actual time required, what is the fair price — answer the question before you see the other side's anchor. The independent number is your defence.
2. When you have already seen the anchor, treat the adjustment as suspicious. If the verdict you arrive at is close to the anchor, the anchor probably did the work. If it is far from the anchor, the adjustment may still be insufficient.
3. Ask what the verdict would be from the other anchor. If the price had opened at half this, what would you have offered? The asymmetry between the two answers reveals the anchor's grip.

Practical steps

1. In negotiation, open first when you have information advantage; let them open when they do. The first number sets the field. The field rarely changes shape afterwards.
2. For consequential numerical decisions, write your independent estimate before exposure to other estimates. The estimates of others, however reasonable, become anchors the moment you see them.
3. Be wary of bracketed displays — the high-anchor decoy, the most-expensive-first layout, the strike-through pricing. Each is an anchor engineered to make a target verdict feel reasonable.
4. For time estimates, double and explain. Anchoring on the deadline produces under-estimates almost universally. The honest estimate often arrives by mentally moving the anchor toward what you would expect of someone else doing the same work.
5. Notice the residue. Where have your numerical verdicts clustered close to anchors that you, on reflection, should not have weighted? The pattern is your own anchoring profile.

Reflection questions

• Where in your life is someone consistently setting the anchor — a salary band, a price list, a deadline culture?
• For one recent verdict, what would the verdict have been if the anchor had been half what it was? Twice?
• Which of your numerical defaults — what a fair price is, what a reasonable salary is, what a normal time is — were set by anchors you did not choose?
• What would change if you generated your own number before every consequential exposure?

Frequently Asked Questions