What a High Divergence Score Means, Even Next to a High Bias Score
A high divergence score means the four data layers we read on a stock, price structure, news sentiment, analyst coverage and fundamentals, disagree with each other on that day. On our 0 to 100 scale, 50 and above counts as high. It does not mean the stock is risky, cheap, expensive or about to move. It means the bias score beside it is a compromise between readings that do not agree, and is best read as one.
The case that confuses people is a high divergence next to a high bias score. If the data disagrees, how does the score reach +88?
A high score can rest on one loud layer
On 17 September 2026, Skyworks Solutions (SWKS) read +88, Broadly Bullish, with divergence 80. Illumina (ILMN) read +88 with divergence 70. Matador Resources (MTDR) read +84 with divergence 58.
The shape behind all three is similar. Price structure sat near the top of its range, and at least one other layer pointed the other way. For Skyworks and Illumina it was fundamentals. For Matador it was news sentiment, which had turned negative in a single session while the other three held.
A bias score is where the layers land on balance. When one of them is extreme it can pull the balance a long way, even with another leaning against it. Divergence is the record that the pull was contested.
Set Elevance Health (ELV) beside them. Same day: +85, divergence 7. One point of bias away from Matador, and an entirely different reading underneath.
High divergence makes the score jumpy
The practical side effect is instability. When the layers split, the balance point sits on a knife edge, and a modest change in one layer swings the composite.
AppLovin (APP) in the second week of September 2026:
- 9 September: bias -8, divergence 51
- 10 September: bias +36, divergence 21
- 11 September: bias +7, divergence 47
- 14 September: bias -13, divergence 62
A 49 point swing in bias over four sessions, from Neutral to Bullish and back down to Mildly Bearish. On the one day the layers briefly lined up, divergence fell to 21 and the score rose. When they split again, it fell back. Anyone watching only the bias score saw a stock changing its mind daily. The divergence column shows why: the disagreement underneath never really went away.
American Tower (AMT) did the same over a longer stretch. Its bias went from +48 on 3 September to -11 on 9 September, while divergence stayed between 52 and 62 the whole time. Analyst coverage held steady and positive. Price structure swung from one side to the other and took the composite with it.
What to take from a high reading
A high divergence tells you the bias score is the least informative number on the page. The score says where the layers balanced; the gap says the balance is not settled. For a stock you follow, the next question is which layer is pulling away, and whether that is new or has been true for weeks.
It says nothing about outcome. Disagreement sometimes closes in the direction price was pointing, sometimes in the direction of the fundamentals, and often it just stays open. Skyworks was at divergence 84 on 18 September, higher than the day before. We do not attach a probability to any resolution, because the data does not support one.
Why it clusters
High divergence is not spread evenly. On 17 September 2026, 16 of the 68 healthcare names we score sat at 50 or above, and the sector averaged 36.1 against 28.5 for the market. Drug developers and medical technology companies often see price react to trial results or approvals long before the reported figures change, which gives price structure and fundamentals a structural reason to disagree. Technology carried 17 names at 50 or more the same day, on a larger base of 107.
The current state of each is on the healthcare and technology divergence boards, and the whole market is on the main board. The general guide to both numbers is what stock divergence means.
Find the layer that is pulling away
The free watchlist report shows all four data layers behind the score for up to five stocks you choose.
Score my five stocks →EquityBias is a market data research tool. Nothing here is financial advice. A high divergence reading describes disagreement in current data and does not indicate future price movement. Readings quoted are historical and dated.
Not investment advice. A Bias Score describes market data that already exists, it does not forecast price, and past readings say nothing about future results. Trading equities carries risk and you can lose part or all of what you put in.
Produced by EquityBias, Jones Inc., CVR DK34325170, Denmark. Responsible: Nicky Jørgensen. Not supervised by a financial authority. Methodology · Interests and conflicts