Control Chart Calculator
The individuals (I-MR) chart from your series in production order: center line, 2.66·MR̄ limits, moving-range limits, and the out-of-control flags — computed from the process, never from a spec.
Control Chart Calculator
Results recalculate instantly on every keystroke. Nothing you type is transmitted.
What this result does not account for
- Individuals (I-MR) form only — no X̄-R subgroup charts, no p/c charts
- One runs rule (8 on a side) — not the full Nelson set
In short: Series 50.2, 49.8, 50.4, 50.1, 49.9, 50.3, 50.0, 50.5, 49.7, 50.2: the center line is x̄ = 50.110000 and the average moving range is MR̄ = 0.444444 over 9 ranges. The individuals limits sit at x̄ ± 2.66·MR̄ = (48.927778, 51.292222); the moving-range chart flags any single jump above 3.267·MR̄ = 1.452000 — the largest here is 0.800000. No point breaches, no 8-point run sits on one side: the chart reads in control. Control limits are the VOICE OF THE PROCESS — they come from these measurements, not from a specification, and that is the whole difference between control and conformance.
Formula
CL = x̄ · I limits = x̄ ± 2.66·MR̄ · MR UCL = 3.267·MR̄
2.66 = 3/d₂ with d₂ = 1.128 for pairs — three sigma estimated from the average moving range of consecutive points. 3.267 is D₄ for subgroups of 2. The constants convert short-run variation into limits; they are estimated FROM this series, which is why a series with a spike grows wider limits — the outlier inflates its own fences, and the flag list says so.
Worked Example
- Enter measurements in the order they were produced — time order IS the chart.
- Read the verdict, then the limits it came from: they belong to the process, not to your spec.
- Check the runs card: a run of 8 on one side of the center line signals a shift before any breach.
- If a point flags, investigate the cause — the chart names suspects, never verdicts.
Defaults: x̄ 50.110000, MR̄ 0.444444, I limits (48.927778, 51.292222), MR UCL 1.452000, no flags — in control. A drive with a 55 spike blows both charts: the point and its moving range flag, and the limits bloat from ±1.18 to ±3.84 — the outlier inflating its own fences, visible on your screen.
Strengths & Limits Of This Model
Where this engine is strong
- Limits estimated from your series; constants shown, not mystified
- Breach + range + runs checks on one panel
Where it stops
- Sensitive to one-off spikes (limits inflate — shown live)
- No phase splits or drifted-limit recalculation
Practical Use Cases
Manufacturing
dimension stability on individuals
Ops
daily counts vs the process' own voice
Teaching
control vs spec, watched live
Methodology & Editorial Standards
Computation runs in IEEE-754 double precision at full internal precision; rounding to two decimal places occurs strictly at the display layer, so no cumulative drift enters the result. All monetary outputs use accounting presentation — grouped thousands, two decimals, negatives in parentheses — so figures can be transcribed directly into a model or working paper. Division-by-zero and out-of-domain inputs return an em-dash rather than a misleading number.
This engine was reconciled against an independent reference implementation and hand-verified for the worked example above before release. Our full five-stage review process is published on the About Us page.
Disclaimer. This calculator is provided for informational and modelling purposes only and does not constitute financial, tax, legal, medical, or engineering advice. Verify all figures with a qualified professional before acting on them.
Control Chart Calculator — 8 Expert FAQs
8 analyst-written answers to the questions practitioners actually ask — optimised for voice and answer-engine retrieval.
Why 2.66? Where does that constant come from?
It is 3/d₂ with d₂ = 1.128, the bias correction for estimating sigma from moving ranges of PAIRS. The 3 is the classic three-sigma control width. Nothing here is typed as a magic number into your limits: 2.66 multiplies the MR̄ computed from YOUR series, so the limits move when the process moves.
Control limits vs specification limits — which is which?
Control limits come FROM the data (this page) and describe what the process DOES; specification limits come from the customer and describe what you WISH. A process can be perfectly in control and hopelessly off-spec — stable is not the same word as capable. The process-capability page prices the spec question; the two pages disagree on purpose.
A point is outside the limits — scrap it?
No: a breach is a SEARCH WARRANT, not a verdict. Investigate what was different about that period — material, setup, measurement — and act on the cause. Deleting the point also shrinks the limits, which quietly makes the chart happier without making the process better.
Why does one spike widen the limits so much?
Because MR̄ averages EVERY consecutive jump, and a spike contributes two huge ranges. The drive in the example note shows it live: one 55 turns ±1.18 limits into ±3.84. That self-inflation is the chart's known weakness with wild one-off events — investigate the spike before trusting the widened limits.
What is the runs rule for?
Eight consecutive points on one side of the center line is an extraordinary streak for a stable process — about a 0.4% event per window. It catches sustained SHIFTS that never quite touch a limit: the process moved and took its limits with it. The runs card counts the longest streak on your series.
Why must the data be in production order?
The moving range is the |difference| between CONSECUTIVE points — shuffle the series and you average unrelated jumps, producing limits that describe nothing. Time order is not a formatting preference; it is half the arithmetic.
My moving range is zero — why the refusal?
A constant series has MR̄ = 0 and the limits collapse onto the center line: the chart would flag every future point as a breach by construction. Zero measured variation means the measurement system may be rounding, or the data are not measurements — either way there is no variation to monitor and the page says so instead of drawing a degenerate chart.
How many points before the limits mean anything?
The page enforces 4 as the arithmetic floor, but estimation honesty wants more: limits from 10 points still wobble. Treat early limits as provisional and re-draw as data accumulate — the verdict card's answer is always about THESE limits, not about the process's soul.