Skip to the invite form

Need help now? Support now: crisis lines by country.

What you are watching

Ask someone in pain how they are, and the answer moves from hour to hour, a little worse and then a little better. One day on its own looks like noise. Look at many days together and a shape appears: a flare that climbs, a peak, a slow way back down.

The M in our logo is drawn the same way. A pen starts low on the page. We flip a coin. Heads, the pen steps up one notch; tails, it steps down one notch. That is all the pen can do. With big notches, a handful of flips gives only a crude zigzag.

Now make every notch half as tall. At fine scales, following the same stretch of time takes about four times as many flips. The zigzag hugs the curve more closely. Keep going, and the steps become too small to see: what is left looks like a single rough line. The mathematician Monroe Donsker proved in 1951 that this always happens. Enough tiny random up-and-down steps, taken quickly enough, draw the jagged kind of curve called Brownian motion, the mathematics of the specks Robert Brown watched jiggling in water.

The M is made of Brownian bridges: random wiggles pinned to pass through a few fixed points (the start, a point on each arch, the dip between them and the end). In the ideal continuous model, reading off the M as smaller and smaller up-or-down moves gives walks that converge to the same path with probability one. The fixed points and the rule that the end pieces stay at or above zero constrain the sequence. Its steps need not be independent, and up and down need not be equally likely.

This belongs on the front door of a pain institute because the scale on the left is the 0 to 10 scale people are asked about every day. The red line, θ, is a flare threshold; the blue line, ρ, is recovery, which is not the same as no pain at all. τθ is the moment pain first reaches the flare line. τρ is the first moment after that when pain is back at recovery. The stretch between them, τρ − τθ, is the time from flare to recovery: what treatment works to shorten. The integral sign stands for adding up a path’s many small steps over that stretch. Read the four letters as words and they spell the Institute’s name: M for Mathematical, ∫ for Institute, ℝ for Research, ℙ for Pain. ℝ and ℙ are the blackboard letters mathematicians use for the real numbers and for probability.

For mathematicians. The M is a concatenation of Brownian bridges (Bessel(3) bridges at the two ends) through five keypoints, seed 60, with volatility σ = 0.42 on the time interval [0, 1], in units where the flare line is θ = 1 and recovery is ρ = 0.15. Here τθ = inf{t : Xt ≥ θ} and τρ = inf{t > τθ : Xt ≤ ρ} are stopping times (first hitting times).

How the path is filled in. The path is stored at 4,097 points, 2−12 apart. Finer detail uses dyadic Brownian-bridge midpoint subdivision (Lévy’s construction). Within each scalar bridge or coordinate bridge, an interval of duration Δ = 2−L with endpoint values a and b receives midpoint (a + b)/2 + σ√(Δ/4) Z. In the ideal stochastic construction, each new midpoint uses an independent standard normal Z, independent also of the existing coarser values. The conditional midpoint law is exactly normal with mean (a + b)/2 and variance σ2Δ/4. Repeated subdivision gives piecewise-linear interpolants that converge almost surely uniformly to the continuous bridge with its prescribed endpoints. The implementation derives pseudorandom normal values from fixed interval-address hashes, giving every visitor the same numerical path and preserving previously generated points. This is a finite-precision numerical realisation of the ideal construction. The end pieces refine their three Brownian-bridge coordinates this way and take the Euclidean norm. Each end piece begins or ends at 0, a foot of the M, so its coordinate bridges have a zero vector endpoint; in the ideal model, independent coordinate bridges then give a Bessel(3) bridge with volatility σ.

How the flips are read and played. Set τ0 = 0 and, after each mark, set τn+1 = inf{t ∈ (τn, 1] : |Xt − Xτn| = h}, with inf ∅ = ∞. Stop when no further move is completed by time 1, and call the number of completed moves N. These marks form the path’s nearest-neighbour hitting-time skeleton. Flip n is H if that move was up and T if it was down, so after n flips with Hn heads and Tn tails, Xτn = X0 + h(Hn − Tn) exactly. The animation plays the N moves at equally spaced times n/N. Their directions inherit the bridge conditioning and need not be independent or equally likely. For the ideal continuous process, [X]t = σ2t. Along dyadic h = 2−k tending to zero, the completed-move count Nh(t) satisfies sup0≤t≤1 |h2Nh(t) − σ2t| → 0 almost surely. Consequently, with N = Nh(1), Nh2 → σ2 and sup0≤s≤1 |τ⌊sN⌋ − s| → 0 almost surely. Linear interpolation of the marked values at times n/N therefore converges uniformly almost surely to X. This is a theorem about refining the same ideal continuous path over the whole interval.

How close. On the original hitting-time clock, the walk’s error over a comparison interval I is DI(h) = supt∈I |walk(t) − path(t)| ≤ 2h. The displayed closeness is the truncated decimal value of 1 − DI(h)/(θ − ρ), using θ − ρ = 0.85 in the original vertical units. For the computed whole-M curve at h = 2−7, with 2,302 moves, the displayed value is 0.983. After 59 nested refinements, the computed value within the final window at h = 2−66 is displayed as 0.999 999 999 999 999 999 974. The latter is a local-window score. Both figures use the original hitting times τn; neither measures error after moving the marks to n/N. They are numerical scores computed using 80-digit decimal working precision and truncated for display.

Mathematical Institute for Research of Pain

…I may have the honour of being your nurse, though I am myself laid low,—though hundreds of miles are between us, and though we can never know one another’s face or voice.

Harriet Martineau, Life in the Sick-Room, 1844She wrote it while ill herself, to a reader in pain she would never meet.

The body is felt by its owner.

Zulu proverb, umzimba uzwiwa ng’umniniwo, in A. T. Bryant, A Zulu-English Dictionary, 1905, p. 729Bryant’s gloss: each one feels his own pain, though unknown to others.

Every one tells me that I look quite blooming and beautiful; and most think I am shamming, but you have never been one of those.

Charles Darwin, letter to J. D. Hooker, 13 June [1850], quoted by Francis Darwin in The Life and Letters of Charles Darwin (printed there as 1849)He was miserably ill that year; his ruddy face perhaps made people think him less ill than he was.

In such extreme accidents, ’tis cruelty to require so exact a composedness. … if the body find itself relieved by complaining let it complain

Michel de Montaigne, Essays, II.37, tr. Charles Cotton, ed. W. C. HazlittHe wrote it living with kidney stones, against the demand to suffer in silence.

Is my strength the strength of stones? or is my flesh of brass?

The Book of Job, 6:12, King James VersionJob, in his affliction: his body is flesh, not stone, and cannot endure without limit.

Ask for an invite

The Institute is open by invitation while it is being tested. Leave your email and, if you like, what kind of pain you live with. A person reads every request and writes back by hand with an invite code, opening the pages that fit what you told us, so it may take a few days.

Privacy for invite requests

What kind of pain do you live with, or work with?

Optional. Tick any that fit.

Your answers come to us as one email, carried by an email service, and this site keeps no copy. We use them only to write to you about your invite (and about testing Nociflux, if you ticked that box), and we delete our copy if you ask.

Have a code? Enter it here.

Need help now? Support now: crisis lines by country.

Paper

Except where a page says otherwise, the Institute’s own text and figures are shared under CC BY-NC 4.0. Pictures and quotations from others keep their own credits and licences.