From "how many survive" to "how risky is right now"
In the previous guide you met the [[survival-function|survival function]] S(x): the probability that a newborn life lives to at least age x, a smooth curve that starts at 1 and slides down toward 0. It answers a stock question — *how many are still here at age x?* But an insurer pricing a policy on a 60-year-old does not really care how many newborns ever reached 60. They care about something sharper: given that this person is alive and 60 *today*, how quickly is death pressing in on them *right now*? That is a flow question, and the survival function alone does not answer it.
The answer is the force of mortality, written with the Greek letter mu and an age subscript — μ at age x. Think of it as the speedometer of death. The survival curve is the odometer: it records how far the journey has gone, how much of the cohort has already died. The force of mortality is the needle showing how fast death is happening at this very instant, for someone who has made it this far. Two ages can sit on a similar part of the survival curve yet face wildly different forces of mortality — just as two cars at the same mile marker can be crawling or flooring it.
What the force of mortality actually measures
Build the idea from the inside out. Take a person alive at exact age x and ask: over the next sliver of time — a tenth of a year, then a hundredth — what fraction of such people die? Crucially this is a [[conditional-probability|conditional probability]]: it is conditioned on *having reached x alive*. We already know the deaths fall only on the survivors, never on those long gone. Divide that small death probability by the small length of time, then squeeze the window toward zero. What you are left with is the death rate per unit time, evaluated for a survivor, at the precise instant x. That limit is the force of mortality.
This is exactly the same construction you have already seen elsewhere, wearing different clothes. It is the survival twin of the [[force-of-interest|force of interest]] from the interest rung: there we squeezed the compounding window to zero to find the instantaneous *growth* rate of money; here we squeeze the time window to zero to find the instantaneous *decay* rate of a life. Interest grows a balance forward; mortality erodes a cohort. Both are continuous-time rates that an effective annual figure only summarises. Recognising μ as the force-of-interest idea turned toward death is one of those moments where the actuarial syllabus suddenly feels like one language, not twenty.
The force of mortality is also exactly what statisticians outside insurance call the hazard rate, and what reliability engineers call the failure rate of a component. The mathematics is identical whether the "life" is a person, a light bulb, a marriage, or a lapsing insurance policy: in all of them we ask how fast the survivors are dropping out, given that they have lasted this long. Later in the ladder you will meet claim frequency and ruin theory; the same hazard idea quietly underpins them too. Learn it once here and you have learned it everywhere.
How the force drives the survival curve
The force of mortality and the survival function are not two separate facts about a life — they are two views of one thing, locked together. The link is the heart of the [[survival-density-hazard-relationship|survival–density–hazard relationship]], and the intuition is clean: the survival curve falls, at each age, in proportion to both how many are still alive *and* how hard the force is pushing on them right then. A small force barely dents the curve; a large force tears chunks out of it. The total fall from birth to age x is the accumulation of all the force endured along the way.
Run that logic to its conclusion and a beautiful result drops out: to survive from birth to age x, a life must escape the force of mortality at *every* instant along the way, and the chance of doing so is the running total of the force, fed through a decay. Pile up (integrate) the force from 0 to x, and the survival probability is the exponential of the negative of that pile. So a constant force gives the [[exponential-distribution|exponential]] survival curve you already know — the memoryless one — and a force that climbs with age bends the curve steeply downward at the top. The shape of the survival curve is nothing more than the autobiography of the force that produced it.
Suppose a constant force mu = 0.02 per year (a toy, flat-hazard life).
S(x) = exp( - mu * x ) survival to age x
S(10) = exp(-0.02 * 10) = exp(-0.20) = 0.8187 ~ 82% reach age 10
S(40) = exp(-0.02 * 40) = exp(-0.80) = 0.4493 ~ 45% reach age 40
1-year death prob from age 40, q = 1 - S(41)/S(40)
= 1 - exp(-0.02) = 0.0198
Note: the rate mu = 0.02 and the probability q = 0.0198 are CLOSE
but NOT equal -- the rate is per-instant, the probability is per-year.The J-shape: the signature of a human life
Real human mortality has a force that is anything but constant. Plot μ against age and you get a famous, lopsided shape often called the bathtub or J-curve. It starts *high* in the first weeks of life — birth and infancy are genuinely dangerous — then plunges to its lifetime low somewhere around age 8 to 11, where a healthy child is about as safe as a person ever gets. Through the teens it ticks up (the so-called accident hump, where risk-taking and road deaths bend the curve), settles through the twenties, and then begins its long, relentless climb.
From roughly age 30 onward, the force of mortality climbs in a strikingly regular way — it very nearly *doubles* every eight years or so. That near-geometric rise is the empirical law (Gompertz's observation) behind most of the [[analytic-mortality-laws|analytic mortality laws]] actuaries fit, and it is why a 70-year-old faces a force many times that of a 40-year-old even though both feel perfectly alive. This steep old-age climb, far more than infant deaths, is what dominates the pricing of a future-lifetime for life insurance and the cost of paying life-expectancy-long pensions.
Why actuaries reach for the force — and where to stay honest
If a [[life-table|life table]] already lists the one-year death probabilities age by age, why bother with a continuous force at all? Three reasons. First, the force is the *smooth* object underneath the table's discrete steps — it lets us value benefits paid at the exact moment of death, or split a one-year probability across fractional ages, without pretending deaths only ever happen on birthdays. Second, it composes cleanly: combine two independent risks (say death and lapse, or two diseases) and their forces simply *add*, which probabilities never do so neatly. Third, it reveals structure — the doubling-every-eight-years pattern is invisible in a column of q-values but jumps out the instant you plot μ.
Hold one honesty above all: a force of mortality is a *model of a population*, never a verdict on a person. The curve says that, among many lives like this one, a certain fraction will die per unit time — it cannot tell you what happens to the individual sitting across the desk, who will live or die exactly once. It is also a snapshot of a particular group at a particular time: the J-shape of a wealthy country today is not the J-shape of a century ago, nor of a different population, and the whole curve has been drifting downward as medicine improves. Pricing a fifty-year policy on yesterday's force, with no allowance for that improvement, quietly assumes the future will copy the past — which is precisely the assumption longevity work exists to challenge.
With the force of mortality in hand, the survival rung has its engine. The next guides turn this continuous picture back into the actuary's everyday workhorse — the life table — counting survivors and deaths age by age, and then layer on the refinements real lives demand: deaths at fractional ages, healthier newly-selected lives, and mortality that keeps improving. But underneath every one of those tables, the quiet driver is always the same needle on the speedometer: μ, the instantaneous risk of dying right now.