The lost socks calculator answers a question that sounds trivial and turns out to be a neat piece of combinatorics: if you start with a drawer of complete pairs and lose some individual socks at random, how many wearable pairs are you actually left with? The answer is always worse than people expect, and the reason is worth understanding, because the same structure shows up whenever the value of a thing depends on it being complete.
Arb Digital publishes free calculators across maths, money and everyday life, and this one is a genuinely good teaching example of linearity of expectation — a technique that makes an apparently horrible counting problem collapse into a single line of algebra. The sock drawer is just the friendliest way to explain it.
What This Lost Socks Calculator Does
Enter how many complete pairs you started with and how many individual socks have gone missing. The calculator returns the expected number of complete pairs still in the drawer, along with how many socks remain in total, how many of those are orphans with no partner, how many pairs were destroyed, and — the more interesting figure — how many socks you would have to lose before the expected pair count falls to half its original value.
Every socks-lost figure it reports is an expected value, not a guarantee. Losing five socks from twenty pairs might destroy as few as three pairs, if the losses happen to come two at a time from the same pairs, or as many as five if every lost sock came from a different pair. The expectation sits between those extremes and is what you would see averaged over many drawers.
The model assumes the socks are lost independently of which pair they belong to, and that every pair is distinguishable — different colours, patterns or sizes, so a surviving sock can only be matched with its own original partner. Both assumptions are discussed further down, because both are worth questioning.
How to Use It
- Count pairs, not socks. The first field wants the number of complete pairs you began with. Twenty pairs is forty socks, and getting this the wrong way round doubles or halves everything downstream.
- Enter how many individual socks are missing. If you know how many pairs are broken rather than how many socks are gone, you already have most of the answer and do not need the calculator.
- Read the hero figure — the expected number of complete pairs left. Compare it against your starting pair count to see the damage.
- Look at the orphan count. Those are the socks taking up drawer space and doing nothing, and they are the reason a half-empty sock drawer still looks full.
- Check the halving figure. It is usually far smaller than people guess, and it is the single most surprising number this tool produces.
The Formula / How It's Calculated
Start with n pairs, which is 2n socks, and lose k of them at random. Rather than trying to count all the ways the losses could be distributed, ask a much simpler question about one specific pair: what is the probability that it survives intact? It survives if neither of its two socks was among the k lost. The number of ways to choose k socks that avoid this pair entirely is the binomial coefficient C(2n−2, k), and the total number of ways to choose any k socks is C(2n, k). The ratio simplifies beautifully:
P(a given pair survives) = (2n − k)(2n − k − 1) ÷ [2n(2n − 1)]
Now the trick. Because expectation is linear, the expected number of surviving pairs is just n times that probability — no matter how the survival of one pair correlates with another. That correlation is real and messy, and linearity of expectation lets you ignore it completely. So:
E[intact pairs] = n × (2n − k)(2n − k − 1) ÷ [2n(2n − 1)]
Worked example, matching the defaults. With n = 20 pairs (40 socks) and k = 5 socks lost: (40 − 5) × (40 − 5 − 1) = 35 × 34 = 1,190, and 40 × 39 = 1,560, so each pair survives with probability 1,190 ÷ 1,560 = 0.762821. Multiply by 20 pairs and the expected number of intact pairs is 15.26. Thirty-five socks remain, of which 2 × 15.26 = 30.51 are in pairs, leaving 4.49 orphans. Losing 12.5% of the socks destroyed 23.7% of the pairs.
Why Pairs Fall Roughly Twice as Fast as Socks
The intuition is simple once you see it. Every sock you lose takes an entire pair out of service, and unless a second loss happens to hit that same already-broken pair, the survivor becomes a useless orphan. So a single loss costs you one pair and gains you one dead sock. Early on, when almost every pair is intact, essentially every loss lands on a fresh pair, and the pair count falls at very nearly one pair per sock — that is, twice as fast in proportional terms as the sock count.
That is why the halving figure is so brutal. With twenty pairs, the expected pair count drops below ten after only twelve socks are lost — twelve out of forty, less than a third of your socks, for half your wearable pairs. Most people guess something closer to twenty. The gap between the guess and the answer is the whole point of the tool.
As losses mount the effect weakens, because more and more of the losses land on pairs that were already broken and cost you nothing further. That is why the curve flattens rather than hitting zero at k = n. The same shape appears in any system where a component failure disables an assembly: the failure rate of assemblies always outruns the failure rate of parts.
The Odds That Every Loss Hits a Different Pair
The worst realistic case is that each of your k lost socks came from a different pair, destroying k pairs. That is not a freak outcome — it is usually the most likely one. The probability is C(n, k) × 2k ÷ C(2n, k): choose which k pairs get hit, choose which sock from each, and divide by all the ways to lose k socks.
With twenty pairs and five losses that comes to 15,504 × 32 ÷ 658,008 = 0.754, so roughly three quarters of the time all five losses are from different pairs and you lose the maximum five pairs. The hero subtitle on this page reports that probability for your own inputs. It is the mathematical version of the feeling that the dryer never eats both socks of the same pair — because with a large drawer and few losses, it very rarely does.
This is essentially a birthday-problem structure in reverse, and the same combinatorics drives our probability calculator and lottery odds calculator for questions that look nothing like laundry.
When the Assumptions Break
Two assumptions carry all the weight, and both are worth checking against your own drawer.
Before either of those, note what the model is at heart: sampling without replacement from a finite population, the same structure formalised by the hypergeometric distribution. Each sock removed changes the odds for the next one, which is exactly why the survival probability has (2n − k) and (2n − k − 1) in it rather than the same term squared.
The first assumption is independence. The model assumes losses are unrelated to which pair a sock belongs to. In practice they often are not: a pair worn together, washed together and lost together is one event, not two. If your socks tend to disappear in twos, the real damage is closer to k ÷ 2 pairs than to k pairs, and this calculator will overstate the harm. If, on the other hand, you lose socks one at a time over months, independence is a decent approximation and the model fits well.
The second is distinguishability. If all your socks are plain black in a single style, they are interchangeable, and the concept of a broken pair barely exists — you have 2n − k socks and can form floor((2n − k) ÷ 2) pairs, with at most one orphan ever. That is the strongest possible argument for buying socks in one style, and it is a real strategy, not a joke. Twelve identical pairs lose one pair per two socks lost; twelve distinguishable pairs lose one pair per sock. Uniformity roughly halves the cost of losing socks, which is why the option in the dropdown points here rather than trying to model a hybrid drawer.
Most real drawers sit between the two extremes, with several groups of identical socks. The honest way to model that is to run this calculator separately for each group of distinguishable styles and add the results, treating each group's internal identical socks with the simple floor rule.
Expected Value Is Not a Prediction
The headline figure of 15.26 pairs is an average across every way five socks could have gone missing. Your drawer contains a whole number of pairs and will show 15, 16 or occasionally 17. The decimal is not a mistake or a rounding artefact — it is the correct answer to a question about averages, and rounding it before you use it in further arithmetic is how small errors compound.
This is the same distinction that catches people out in gambling and insurance maths, where an expected value describes the long run and says almost nothing about a single trial. Our expected value calculator covers the general case, and the coin flip probability calculator shows how wide the spread around an expectation can be for small sample sizes.
Where Else This Structure Turns Up
The sock problem is a disguised version of a common engineering question: given a system built from paired or grouped components, how many complete assemblies survive a random loss of parts? The answer always has the same shape, and the same linearity-of-expectation shortcut applies — compute the survival probability for one assembly, multiply by the number of assemblies, and stop.
You see it in spare-parts planning, in matched sets of anything sold as a set, and in redundancy analysis where a subsystem needs all of its members present. The reason the sock version is memorable is that everyone has personally experienced the drawer full of orphans, which is exactly what the arithmetic predicts. For related counting problems, the dice probability calculator and conditional probability calculator use the same toolkit on different questions.
Arb Digital maintains a large library of free tools covering maths, money, home and work — no signup, no limits. Have a look, or get in touch if there is one you wish existed.
Browse Free Tools Contact Arb DigitalCommon Mistakes to Avoid
- Entering socks where the field asks for pairs — the first input is complete pairs, and doubling it silently doubles every result.
- Reading the expected value as a certainty — 15.26 pairs is an average over many drawers, not a prediction about yours.
- Assuming losing k socks destroys exactly k pairs — it is the most likely outcome for small k, but not the only one, and it stops being true once losses get large.
- Applying this to a drawer of identical socks — interchangeable socks follow a completely different and much gentler rule, roughly one pair lost per two socks.
- Ignoring the correlation when socks disappear in twos — if a pair vanishes together, that is one event, and the independence assumption behind this model no longer holds.
Related Free Tools From Arb Digital
For the general case, use the probability calculator or the expected value calculator. The coin flip probability calculator and dice probability calculator handle the classic independent-trial questions, the conditional probability calculator covers dependence directly, and the lottery odds calculator applies the same binomial coefficients at a much larger scale. Everything else is in the free online tools hub.
Frequently Asked Questions
Because it is an expected value — an average over every possible way the losses could have fallen. Your actual drawer will hold a whole number of pairs, but the average across many identical scenarios lands between whole numbers, and rounding it before doing further arithmetic introduces error.
No, but with a large drawer and few losses it usually does. If two of the lost socks happen to come from the same pair, only four pairs are affected. The calculator reports the probability that every loss hit a different pair, which for small loss counts is typically the most likely outcome.
The whole problem changes. Interchangeable socks can be paired with any other sock, so you simply halve the remaining count and round down, leaving at most one orphan ever. Buying socks in a single style genuinely does roughly halve the cost of losing them.
It is the rule that the expected value of a sum equals the sum of the expected values, even when the terms are not independent. It lets you compute the survival probability of one pair and multiply by the number of pairs, without ever untangling how one pair's fate relates to another's.
Far fewer than most people guess. With twenty pairs, the expected pair count drops below ten after about twelve socks go missing — less than a third of the socks for half the pairs. The calculator reports this figure for whatever drawer size you enter.
It models socks lost independently of each other. That fits losses spread over time reasonably well. It fits less well if a pair tends to vanish together, in which case the true pair damage is closer to half what this tool reports.
Socks whose partner has been lost. They are still in the drawer, still taking up space, and cannot be worn as a pair with anything. They are the reason a drawer can look reasonably full while offering very few wearable options.
This tool is a probability model for entertainment and teaching. It reports expected values under stated assumptions about how socks are lost, and those assumptions may not match your own laundry.
