The Math of Soulmates: 13,400× Harder Than the Lottery — or Basically Inevitable?

A friend tells you: “Meeting you was fate.” What number does that sentence translate to, mathematically?

Roughly 109 billion people have been born and died on Earth before us.^[1] If the universe had pre-selected exactly one person from that entire catalog to be your partner, what are the odds you’d ever cross paths with them? Let’s work that out.

But here’s the thing: once we run the numbers, the calculation splits into two answers that point in opposite directions. One is a number so small it borders on cruel. The other is almost offensively encouraging. Which is more comforting is, genuinely, up to you.

This article is a mathematical thought experiment. It does not argue for or against any particular worldview, religion, or theory of fate. It does not claim to settle whether destiny “exists.” The assumptions drive the conclusions — so scrutinize the assumptions more than the answers.

INPUT

Branch 1 — Historical perspective: the odds of meeting one specific person across all of human history

Variable 1: Total deceased humans $N_{\text{dead}}$

$$N_{\text{dead}} = 1.09 \times 10^{11} \text{ people}$$

According to the Population Reference Bureau (PRB)'s 2022 updated estimate, approximately 117 billion humans have been born on Earth in total.^[1] Subtract the roughly 8.2 billion currently alive (UN DESA, 2024)^[2] and you get approximately 109 billion who have already died. The PRB counts from around 190,000 BCE, starting with a world population of 2 — a conservative baseline relative to the ~300,000-year anatomical history of Homo sapiens, but the ultra-low population density of that early period makes the difference negligible in the cumulative total.

This article and the how-big-is-heaven article use the same PRB source and the same figure.

Stance: central estimate.

Branch 2 — Realistic perspective: expected compatible people within a lifetime of encounters

Variable 2: Accessible dating pool $N_{\text{pool}}$

$$N_{\text{pool}} = N_{\text{world}} \times f_{\text{gender}} \times f_{\text{age}} \times f_{\text{geo}}$$

Central values for each filter:

• $N_{\text{world}} = 8.2 \times 10^9$: UN DESA World Population Prospects 2024.^[2]
• $f_{\text{gender}} = 0.50$: Fraction of opposite-gender population. Based on the UN’s global birth sex ratio of ~105:100. Adjust to taste for same-sex attraction.
• $f_{\text{age}} = 0.25$: Share of the world population within ±10 years of your own age. The 25–45 bracket accounts for roughly 25% of total world population in UN 2024 data.^[2] A tighter ±5-year window would drop this to ~13%. Stance: central (±10 years).
• $f_{\text{geo}} = 0.006$: Geographic accessibility. Anchored to South Korea’s population (~52 million) divided by world population (~8.2 billion) ≈ 0.006, representing “domestic Korean opposite-gender adults” as the baseline.^[3] For English-speaking readers, the US equivalent would be roughly 0.041 (330 million / 8.2 billion) — about five times wider a slice. Expanding to East Asia raises this to ~0.30; using fully global online dating pushes it to 1.0.

$$N_{\text{pool}} \approx 8.2 \times 10^9 \times 0.50 \times 0.25 \times 0.006 \approx 6.2 \times 10^6 \text{ people}$$

Stance: domestic-Korea baseline, conservative-to-central. Adjustable via the slider below.

Variable 3: People you actually meet in a lifetime $N_{\text{meet}}$

$$N_{\text{meet}} = 10{,}000 \text{ (central estimate)}$$

This is the second-most sensitive assumption in the entire calculation. The definition of “meeting” changes the figure by an order of magnitude or more.

• 150 people: Dunbar’s number. British anthropologist Robin Dunbar (1992) derived this upper limit on stable human social groups from the correlation between neocortex size and group size in primates.^[4] Hill & Dunbar (2003) confirmed it empirically with a measured mean of 153.5.^[5] This is the ceiling on “people you genuinely know.” Stance: conservative lower bound.
• 1,500 people: Face-and-name acquaintances — colleagues, classmates, neighbors, summed across a life.
• 10,000 people: Anyone you’ve exchanged more than a few words with, at least once. The sociological central estimate and this article’s default. Stance: central.
• 80,000 people: Broad “contacts” including social media, dating apps, and online communities. Stance: aggressive upper bound.

Variable 4: Compatibility filter $f_{\text{compat}}$

$$f_{\text{compat}} = 0.01 \text{ (central; 1 in 100)}$$

This is simultaneously the most sensitive variable in the model and the one with the weakest academic grounding. What fraction of people of the right gender and age would you actually want to be with — factoring in appearance, personality, values, lifestyle, humor — has no agreed empirical value in psychology.^[6] This article uses 1% (1 in 100) as a placeholder central estimate and provides a slider so you can substitute your own number. The final result scales linearly with this value, which is precisely why plugging in your own figure is the point.

Stance: central, no strong academic basis — reader input strongly encouraged.

Variable 5: Availability (single) rate $f_{\text{single}}$

$$f_{\text{single}} = 0.50$$

South Korea’s Statistics Korea 2023 marriage data shows that among adults aged 20–44, the share who are unmarried or not currently in a committed relationship varies by age group, but a rough aggregate estimate across the eligible adult population comes to about half.^[3] This figure shifts with time and geography. Stance: central.

FORMULA

Branch 1: Historical probability — the odds of meeting one specific person

Assume “the one” is exactly one person, and that person could be any of the ~109 billion people who have ever lived. The prior probability $P_{\text{hist}}$ that a given individual is your fated partner is:

$$P_{\text{hist}} = \frac{1}{N_{\text{dead}}} = \frac{1}{1.09 \times 10^{11}}$$

Which gives:

$$P_{\text{hist}} \approx 9.17 \times 10^{-12}$$

Let’s compare this to something familiar. The Korean Dong-haeng lottery (6/45) — pick 6 numbers out of 45, order doesn’t matter — has a jackpot probability of:^[7]

$$P_{\text{lotto}} = \frac{1}{\binom{45}{6}} = \frac{1}{8{,}145{,}060} \approx 1.23 \times 10^{-7}$$

For reference, the US Powerball jackpot sits at roughly $\frac{1}{292{,}201{,}338} \approx 3.4 \times 10^{-9}$ — which itself is still thousands of times more likely than $P_{\text{hist}}$.

The ratio:

$$\frac{P_{\text{hist}}}{P_{\text{lotto}}} = \frac{9.17 \times 10^{-12}}{1.23 \times 10^{-7}} \approx \frac{1}{13{,}400}$$

$P_{\text{hist}}$ is about 13,400 times smaller than the odds of hitting the Korean lottery jackpot. Winning the lottery is already a stretch; “the one specific pre-selected person across all of human history” is a stretch’s stretch.

This calculation rests on a hidden premise: that “the one” could be literally anyone ever born. Adjust that premise even slightly and the entire story changes.

Branch 2: Realistic expected count — people you’ll actually meet in a lifetime

Step 1: Calculate the accessible pool

$$N_{\text{pool}} = N_{\text{world}} \times f_{\text{gender}} \times f_{\text{age}} \times f_{\text{geo}}$$

$$N_{\text{pool}} = (8.2 \times 10^9) \times 0.50 \times 0.25 \times 0.006$$

$$N_{\text{pool}} = 8.2 \times 10^9 \times 7.5 \times 10^{-4}$$

$$N_{\text{pool}} \approx 6.15 \times 10^6 \approx 6.2 \text{ million people}$$

On the domestic-Korea baseline, about 6.2 million people are of the right gender and in your age range — roughly two-thirds of Seoul’s population (~9.5 million).

Step 2: Compatible candidates within that pool

$$N_{\text{candidate}} = N_{\text{pool}} \times f_{\text{single}} \times f_{\text{compat}}$$

$$N_{\text{candidate}} = 6.15 \times 10^6 \times 0.50 \times 0.01$$

$$N_{\text{candidate}} \approx 3.075 \times 10^4 \approx 30{,}000 \text{ people}$$

At 1% compatibility and 50% availability, roughly 30,000 people in South Korea alone qualify as potential partners on paper.

Step 3: Expected lifetime encounters with those candidates

Assuming $N_{\text{meet}}$ already incorporates the gender/age/geography filters (i.e., it represents contacts within your realistic social sphere), the simplified expected count is:

$$N_{\text{expect}} = N_{\text{meet}} \times f_{\text{compat}} \times f_{\text{single}}$$

$$N_{\text{expect}} = 10{,}000 \times 0.01 \times 0.50$$

$$\boxed{N_{\text{expect}} = 50 \text{ people}}$$

Under central assumptions (domestic baseline, 10,000 lifetime encounters, 1% compatibility, 50% availability), you can expect to meet roughly 50 potential partners over the course of your life.

Dimension a: Expected count within Dunbar’s 150 — among people you truly know

What if “the one” only counts if they come from your intimate social circle?

$$N_{\text{dunbar}} = 150 \times f_{\text{compat}} \times f_{\text{single}}$$

$$N_{\text{dunbar}} = 150 \times 0.01 \times 0.50 = 0.75 \text{ people}$$

The expected number of compatible, available partners within your closest 150 relationships is 0.75 — meaning three out of four people will not find a match in that inner circle. Dunbar (1992) derived this upper bound from neocortex ratios^[4] and Hill & Dunbar (2003) confirmed it empirically.^[5] The ceiling on how many people we can truly know is a genuine anthropological constraint. If fate is limited to “people you really know,” you may simply need to get lucky.

Dimension b: What online dating actually changes

Stanford’s “How Couples Meet and Stay Together” study (Rosenfeld, Thomas, and Hausen) found that the share of US couples who first met online rose from ~22% in 2009 to ~39% by 2017.^[8] The trend in South Korea mirrors this pattern closely, with dating app usage growing sharply.^[9] In practical terms, $f_{\text{geo}}$ can now realistically expand from “domestic” (0.006) to “East Asia” (0.30) or “global” (1.0).

Here is the counterintuitive result. Widening $f_{\text{geo}}$ to global balloons $N_{\text{pool}}$ from 6.2 million to over 1 billion — but $N_{\text{expect}}$ is capped by $N_{\text{meet}}$. No matter how large the pool gets, the bottleneck is the number of people you can actually spend meaningful time with in one lifetime.

$$N_{\text{expect,global}} = 10{,}000 \times 0.01 \times 0.50 = 50 \text{ people}$$

The expected count is identical. What online dating changes is not “how many compatible people you’ll meet” but “how many compatible people exist somewhere that you will never meet” ($N_{\text{candidate}}$). A bigger pool means more missed chances, not more guaranteed encounters.

Dimension c: Reverse calculation — how many compatible people exist in the world?

Remove $f_{\text{geo}}$ entirely and ask: globally, how many people would be compatible with you?

$$N_{\text{soulmate\_world}} = N_{\text{world}} \times f_{\text{gender}} \times f_{\text{age}} \times f_{\text{single}} \times f_{\text{compat}}$$

$$N_{\text{soulmate\_world}} = (8.2 \times 10^9) \times 0.50 \times 0.25 \times 0.50 \times 0.01$$

$$N_{\text{soulmate\_world}} \approx 5.125 \times 10^6 \approx 5.1 \text{ million people}$$

Somewhere on Earth, roughly 5.1 million people would be compatible with you. That is more than half of New York City’s population, or about 10% of South Korea’s entire population.

“The one” may not be one. The problem is not probability. It is radius of opportunity.

Scenario comparison table

Fixed: $N_{\text{meet}} = 10{,}000$, $f_{\text{gender}} = 0.50$, $f_{\text{single}} = 0.50$.

Post-hoc confirmation: the probability of meeting at least one

Branch 1 and Branch 2 give radically different answers because they answer different questions.

• Branch 1’s question: “What is the probability that a specific pre-assigned person is my soulmate?” → Essentially zero.
• Branch 2’s question: “What is the probability that I’ll meet at least one person who meets my criteria?”

The second question is a standard binomial probability. Over $N_{\text{meet}}$ encounters, where each encounter has probability $p = f_{\text{compat}} \times f_{\text{single}}$ of producing a compatible, available person:

$$P(\text{at least one}) = 1 - (1 - p)^{N_{\text{meet}}}$$

Plugging in central values: $p = 0.01 \times 0.50 = 0.005$, $N_{\text{meet}} = 10{,}000$

$$P(\text{at least one}) = 1 - (1 - 0.005)^{10{,}000} = 1 - 0.995^{10{,}000}$$

Exponential approximation ($e^{-np}$ rule): $(1-0.005)^{10{,}000} \approx e^{-0.005 \times 10{,}000} = e^{-50} \approx 1.93 \times 10^{-22}$

$$P(\text{at least one}) \approx 1 - 10^{-22} \approx 1.000$$

Near-certainty. Which means the feeling of “this person was meant for me” is not the result of astronomical rarity. Meeting someone who clears your criteria over a lifetime is close to inevitable. What actually happens is that we meet someone, choose them, and then — only afterward — call it fate. That is post-hoc confirmation: the label gets written after the fact.

OUTPUT

Branch 1 is blunt. The probability that a single pre-assigned person out of all 109 billion who ever lived is your partner: roughly $9.2 \times 10^{-12}$, about 13,400 times smaller than the odds of winning the Korean lottery jackpot — or, to put it in terms an American reader will feel in their gut, roughly 3,500 times smaller than hitting Powerball. The scenario where the universe bookmarked exactly one person for you is, mathematically, the most romantic and simultaneously the most punishing setup imaginable.

Branch 2 runs in the opposite direction. Reframe the question from “that specific one person” to “anyone who meets my criteria,” and the central-case expected count over a lifetime is roughly 50 people. Somewhere in the world right now, approximately 5.1 million people would be compatible with you. And the probability of meeting at least one compatible, available person in your lifetime, at central assumptions, is 1.000 — with the rounding error hiding 22 decimal places down.

The Dunbar finding is worth a pause: within your closest 150 relationships, the expected number of compatible, available partners (at 1% compatibility) is 0.75. Waiting for fate to land in your intimate social circle is a strategy that statistically fails three in four people.

Set those three results side by side and one conclusion comes into focus. The prior probability that any specific person was pre-destined for you approaches zero. The posterior probability — once you’ve met someone and chosen them — that someone fitting your criteria would turn up in your lifetime is near-certain. Fate may not be something discovered. It may be a label we apply after the choosing.

Whether any of this means fate “exists” is not a question this calculation answers. That part belongs to you. What the math does say is this: with 5.1 million compatible people in the world, the obstacle is not probability. It is radius.