Same skill. Same career-defining events. Only the order they arrived in is random — and that alone splits a #3 finish from a #847 finish.
Every "career" simulated below starts with identical true skill (the selection-strength dial) and draws
from the exact same shared pool of variance events — bad beats, hero calls, lucky bracket seeds, rough draws.
The only thing that differs between careers is the shuffle: the random order those events happen to arrive in.
Early swings compound for longer, so an identical event landing early founds a much bigger "clone" of success than
the same event landing late — same math as a tumor: a driver mutation that lands early gets more generations to
expand a clone; the same mutation landing late doesn't. Cranking selection strength to 100% makes every career
literally identical (spike). Dropping it to 0% turns the whole leaderboard into pure order noise, with real skill
contributing nothing. Before you crown someone "the best," ask: is he actually good, or did he just run good?
Model: arXiv:2609.25838 — "Mutation Order and Selection Shape Intratumor Heterogeneity in Tumor Evolution".
Controls
How many identical-skill "players" run the gauntlet
0% = pure order noise · 100% = skill fully determines outcome
Size of the shared event pool every career experiences
Best / worst ratio
—
Spread (log₁₀ std dev)
—
Median career score
—
Careers run
—
Outcome distribution
Every bar/dot below comes from the exact same skill level and the exact same set of events — only the arrival order differs.
Leaderboard — identical skill, wildly different fates