Why your entry ownership percentage is almost irrelevant to your return
A GP negotiates a Seed check and walks away with 8% ownership. That number is written in the term sheet, noted in the investment memo, and then systematically eroded over the next 5–8 years by every subsequent financing round. The cap table tracks each step. But no cap table tool tells you what the distribution of that erosion looks like across all plausible future paths — and that distribution is what determines your return.
The question that matters is not "what do I own today?" It is: "what is the distribution of what I will own at exit?" That question requires modelling every future round — its size, valuation, and dilution — not as a fixed assumption but as a probability distribution calibrated to real deal data. The difference between those two approaches is the difference between a cap table and a dilution model.
Dilution compounds. A 20% dilution at Series A, followed by 18% at Series B, followed by 15% at Series C, leaves you with 56% of your original position — not the arithmetic average. Most back-of-envelope dilution estimates get this wrong by adding rather than multiplying. The correct model is multiplicative at every stage, applied to a probability distribution of round sizes and terms.
The three uncertain inputs that determine your dilution at every future round
A complete dilution model requires three inputs per future financing stage:
- Round size ($M raised) — calibrated to stage benchmarks. Seed rounds typically raise $2–8M; Series A $8–25M; Series B $25–75M. Each has a distribution based on vertical and year.
- Pre-money valuation — determines the price per share and your resulting dilution percentage. Valuation benchmarks vary significantly by stage, vertical, and growth trajectory. A P75 CGI company commands a higher Series A pre-money than a P25 company.
- Probability of reaching this stage — not every Seed company raises a Series A. The Pv × Pc survival model provides the stage-conditional survival probability, which weights the ownership path by whether the company actually reaches each round.
Why 10,000 simulations replace a single dilution estimate
Running 10,000 simulations through all financing stages produces the full ownership distribution at exit. Each simulation draws a random round size, valuation, and dilution for each stage from calibrated distributions based on real deal data for the vertical. The output is a P10/P50/P90 ownership band at exit — not a single number, but the range of plausible outcomes weighted by their probability.
The P50 ownership path is your base case: what you'd expect at median round sizes and valuations for the vertical. The P10 is the heavy-dilution scenario — large rounds at low valuations, down rounds, flat rounds. The P90 is the efficient capital path — the company raises at premium valuations with minimal dilution because its growth trajectory commands it. That spread between P10 and P90 is the uncertainty that most GPs ignore when they model dilution with a single number.
The dilution model interacts directly with your pro-rata strategy. If you hold pro-rata rights and exercise them at every stage, your dilution path is different — and more expensive in capital terms. The question is whether the ownership preservation is worth the additional capital deployed. That is a portfolio-level capital allocation decision (Module 6), not a deal-level one.
How growth trajectory determines the valuation you'll face at each future round
Future round valuations are uncertain — but they are not random. They follow predictable distributions calibrated to stage and vertical data, with one important adjustment: a company's CGI growth percentile is the strongest predictor of the valuation premium it commands at each future stage. A company tracking at P75 in its cohort commands meaningfully higher Series A and B pre-money valuations than a P40 company in the same cohort — and that difference compounds across every subsequent round.
The framework calibrates round valuation distributions to real deal data by stage and vertical, then adjusts for CGI percentile. A P75 company gets a valuation premium applied to its distribution inputs — producing a narrower, higher-centred P10/P90 ownership band than a P25 company facing less favourable round terms at every stage.
The complete picture: combining dilution and exit distributions to get true investor return
Combining the dilution model with the exit value probability model gives you the only number that actually matters for fund construction: investor value at exit = exit value × ownership at exit. Both are distributions. Their product — computed via Monte Carlo — produces a full P10/P50/P90 distribution of investor return. That distribution is what drives check size decisions, reserve allocation, and pro-rata strategy. Everything else is a simplification of it.