Trace one lateral cephalogram three ways — Steiner, Downs, and Ricketts — and you can come away with three subtly different impressions of the same face. One analysis calls the skeletal pattern borderline Class II; another, reading the profile, leans toward a balanced convexity; a third, weighing the growth vector, hints the mandible will catch up. None of that is a tracing error. It is what happens when three systems, built in different decades from different reference planes, describe the same jaws in their own vocabulary.
This is not a theoretical worry. When investigators applied seven common analyses to the same 125 cephalograms, agreement on the sagittal jaw relationship was only moderate, and agreement on growth pattern and incisor position was merely fair.9 The lesson is not that one analysis is right and the others wrong — it is that each answers a slightly different question, and knowing which question you are asking is half of choosing the right tool.
One clarification before we compare them, because it is the source of most confusion. These analyses do not each stamp a Roman-numeral class on the film. Steiner's ANB angle reads the skeletal class more or less directly. Downs and Ricketts infer the same anteroposterior (AP) relationship indirectly — Downs through the angle of convexity and the facial plane, Ricketts through facial depth and the shape of the profile. So when the "verdicts" diverge, it is three reads on one underlying relationship, filtered through three different measurements — not three independent diagnoses.
The Three Analyses at a Glance
All three are lateral cephalometric analyses, and all three share most of the same landmarks. What separates them is what they were built to do and, crucially, which reference plane they measure from.
Downs came first, described at the University of Illinois in 1948 as a way to characterise the "typical" and "atypical" face using Frankfort horizontal as the baseline.3 Steiner followed in the 1950s, distilling cephalometrics into a compact, teachable set of angles measured from the Sella–Nasion line.12 Ricketts then expanded the field toward growth prediction and computerised analysis, layering several planes to describe not just where the jaws are but where they are heading.4
| Analysis | Primary reference | Built to answer | Signature measurements |
|---|---|---|---|
| Steiner (1953) | Sella–Nasion (SN) plane | Fast anteroposterior and dental screening | SNA, SNB, ANB; U1–NA, L1–NB; interincisal angle |
| Downs (1948) | Frankfort horizontal | Skeletal and profile "type," vertical growth | Facial angle, angle of convexity, Y-axis, mandibular plane angle |
| Ricketts (1960s) | Multiple planes (FH, facial plane, etc.) | Growth direction, vertical control, esthetics | Facial axis, facial depth, mandibular plane, convexity, E-line |
Shared landmarks, different lenses. Because Sella, Nasion, points A and B, Porion, Orbitale, Gnathion, and the incisors serve all three analyses, one careful tracing feeds all of them. The differences you see between the reports come from the reference planes and formulas, not from separate tracings.
Reference Planes Decide Everything
The single most important difference between these analyses is quiet and easy to overlook: Steiner measures from the cranial base, Downs and Ricketts measure from Frankfort horizontal. That choice propagates into every skeletal number the analysis produces.
The SN plane connects Sella to Nasion and approximates the anterior cranial base. It is convenient — both points are reliable and easy to locate — and it is stable enough to superimpose serial films on. But it is not a fixed horizon. When a patient's cranial base is unusually steep or flat, the whole SN line tips, and every angle referenced to it tips with it. That is why a "normal-looking" SNA of 82° can coexist with a genuinely retrognathic maxilla: the number is normal, but the baseline it was measured from is atypical.8
ANB inherits this problem twice over. Because it is the difference between two SN-referenced angles, it is sensitive not only to cranial base inclination but also to the vertical position of Nasion and to jaw rotation.78 This is precisely why the Wits appraisal was introduced — it re-expresses the AP jaw relationship on the functional occlusal plane, sidestepping the cranial base entirely.6
Frankfort horizontal (Porion to Orbitale) is the baseline Downs and Ricketts use, and it is anatomically closer to the natural head posture a clinician sees in the chair. Its weakness is at the other end: Porion and Orbitale are among the harder landmarks to identify consistently, and landmark identification — not measurement — is the dominant source of cephalometric error.5 So the trade is real. SN buys you reproducible landmarks at the cost of a baseline that moves with the cranial base; Frankfort buys you a more clinically meaningful horizon at the cost of shakier landmarks.
Steiner — When to Reach for It, Where It Misleads
Steiner is the analysis almost every clinician learns first, and for good reason. In three angles — SNA, SNB, ANB — it delivers a fast, legible read on the AP position of each jaw and their relationship to each other, then adds incisor positions relative to those same lines.1 It is quick to trace, quick to teach, and quick to communicate. For a routine new-patient screen where you mainly want to know "how far off is this bite, and in which direction," it is hard to beat.
Steiner himself was careful about what the numbers meant. He derived his norms from a specific, fairly small sample and presented them as clinically useful reference points, not universal biological constants.2 That nuance is easy to lose once "SNA 82°, ANB 2°" has been repeated in enough textbooks.
Where it misleads
The blind spot is the one already described: everything hangs on the SN line. In a patient with an atypical cranial base, SNA and SNB drift, and ANB — the headline number — drifts with them. ANB is also affected by facial height and by rotation of the jaws, so two patients with identical ANB values can have visibly different profiles.7 The fix is not to abandon Steiner but to cross-check it: read ANB alongside the Wits appraisal, and when the cranial base looks unusual, confirm the skeletal picture with a Frankfort-referenced analysis.
Downs — When to Reach for It, Where It Misleads
Downs is the oldest of the three and, in a sense, the most profile-minded.3 Because it is anchored to Frankfort horizontal rather than the cranial base, its skeletal readings do not inherit the SN-inclination problem. Two of its measurements are quietly excellent. The angle of convexity summarises the AP harmony of the profile in a single number that is independent of the cranial base. The Y-axis (the Sella–Gnathion line to Frankfort) captures the downward-and-forward direction of chin growth, making Downs genuinely useful for reading the vertical pattern and the growth tendency.
Reach for Downs when the question is about the profile and the vertical dimension rather than a fast AP class — an open-bite tendency, a question about how convex a face really is, or a case where the cranial base makes you distrust the Steiner numbers.
Where it misleads
Two caveats. First, the Frankfort landmarks: if your Porion or Orbitale is off, the facial angle and Y-axis move with it, so the payoff depends on disciplined tracing.5 Second, the norms are old and were drawn from a narrow sample, so treat them as reference ranges rather than targets — a point that applies to all three of these analyses, but bites hardest on the oldest one.
Ricketts — When to Reach for It, Where It Misleads
Ricketts is the most ambitious of the three. Rather than a compact screen, it is a layered description of the face built around growth: the facial axis for the direction of chin growth and vertical control, facial depth for the AP position of the mandible, the mandibular plane angle, convexity of point A, and the esthetic (E-) line for the lips.4 Its native habitat is the growing patient and the vertical or surgical case, where knowing the direction of growth changes the plan.
If you are trying to anticipate how a child's face will change, or to control vertical dimension in a hyperdivergent case, Ricketts gives you measurements the other two simply do not — a growth vector and an esthetic reference in the same pass. Pairing its facial axis with a soft-tissue read makes it a natural fit for cases where the profile and the growth pattern, not just the current class, drive the decision.
Where it misleads
The cost of that richness is error accumulation. More measurements mean more landmarks, and because landmark identification is the leading source of cephalometric variability, a fuller analysis has more places for small errors to enter.5 The complexity can also invite over-reading — treating a growth forecast as a fixed outcome. Used well, Ricketts is powerful; used carelessly, it offers more numbers to misinterpret than Steiner does.
Which Analysis for Which Case
Here is the practical part. Rather than crown one analysis, match the tool to the clinical question in front of you.
Routine Class I/II/III triage and incisor position. Fast and legible — pair with Wits as a cross-check.
Growing patients, hyperdivergent and open-bite cases, surgical planning where growth direction matters.
Profile harmony and vertical pattern, and a Frankfort-referenced cross-check when the cranial base is atypical.
Read as a decision framework: reach for Steiner when you want a quick, communicable AP and dental snapshot; add the Wits appraisal whenever ANB and the profile seem to disagree. Reach for Ricketts when growth prediction or vertical control is the point — a mixed-dentition patient, a hyperdivergent open bite, a borderline surgical case. Reach for Downs when the profile's convexity and the vertical pattern matter, or when an unusual cranial base makes you want a Frankfort-referenced second opinion. And in the classic scenario where a normal SNA hides a retrognathic maxilla, deliberately step off the SN line — Wits, Downs convexity, and the Ricketts facial plane will each expose what the cranial-base-referenced number smoothed over.
A word on how to hold the disagreement between analyses. It is tempting to conclude that if the numbers conflict, at least one must be broken. The more honest reading is subtler. When the same films were analysed by automated tools, the raw values agreed well for most parameters, with the discrepancies concentrated in exactly the places where the analyses define a measurement differently — the angle of convexity and the occlusal plane.10 Yet the same underlying values can still yield different classifications, which is what the seven-analysis agreement study found.9 In other words: a Steiner reading and a Ricketts reading can measure a face almost identically and still hand you different verdicts, because the verdict lives in the threshold, not the millimetre. That is an argument for triangulating, not for trusting one.
The Practical Answer: Run Them Together
The good news is that choosing between these analyses is often a false dilemma. Since they share almost all their landmarks, the marginal cost of a second analysis is nearly zero — the expensive part is placing the points, and you only do that once. The disciplined workflow is to trace carefully, run Steiner for the AP and dental snapshot, and then bring in Downs or Ricketts for the profile and growth context before committing to a plan.
That is the model BCeph is built around: place your landmarks once, and it computes Steiner, Downs, Ricketts, Wits, and ten other analyses from the same tracing, side by side, in your browser. Cross-checking ANB against Wits, or a Steiner class against the Ricketts facial axis, costs no extra tracing time — which is exactly what makes triangulation practical rather than aspirational. When the analyses agree, you can act with more confidence; when they disagree, you know precisely where to look.
Frequently Asked Questions
References
- Steiner CC. Cephalometrics for you and me. Am J Orthod. 1953;39(10):729–755. doi:10.1016/0002-9416(53)90082-7
- Steiner CC. Cephalometrics in clinical practice. Angle Orthod. 1959;29(1):8–29. Full text
- Downs WB. Variations in facial relationships: their significance in treatment and prognosis. Am J Orthod. 1948;34(10):812–840. PubMed
- Ricketts RM. Perspectives in the clinical application of cephalometrics: the first fifty years. Angle Orthod. 1981;51(2):115–150. PubMed
- Baumrind S, Frantz RC. The reliability of head film measurements. 1. Landmark identification. Am J Orthod. 1971;60(2):111–127. doi:10.1016/0002-9416(71)90028-5
- Jacobson A. The “Wits” appraisal of jaw disharmony. Am J Orthod. 1975;67(2):125–138. doi:10.1016/0002-9416(75)90065-2
- Hussels W, Nanda RS. Analysis of factors affecting angle ANB. Am J Orthod. 1984;85(5):411–423. PubMed
- Järvinen S. An analysis of the variation of the ANB angle: a statistical appraisal. Am J Orthod. 1985;87(2):144–146. doi:10.1016/0002-9416(85)90024-7
- Monisha J, Sangeetha U, Nivethitha B, Madhan B. Agreement between cephalometric analyses in diagnosing the dento-skeletal characteristics of malocclusion. J Oral Biol Craniofac Res. 2025;15(4):744–748. doi:10.1016/j.jobcr.2025.04.012
- Kazimierczak W, Gawin G, Janiszewska-Olszowska J, et al. Comparison of three commercially available, AI-driven cephalometric analysis tools in orthodontics. J Clin Med. 2024;13(13):3733. doi:10.3390/jcm13133733
Links open the journal record, PubMed entry, or DOI resolver. Normative values discussed here are documented reference ranges, not universal constants — interpret them against the individual patient and reference plane.