What “the edge” means here
Skill advantage is not a fixed quantity that applies the same way everywhere. It is better described as a tilt in the odds, and the size of that tilt depends on how much room the game leaves for it. At the boundaries, meaning very high randomness or very closely matched players, the tilt gets small enough that it is hard to see and easy to doubt.
This article is about those boundaries. It is not about the everyday mix of the two forces, but about the points where one of them nearly drowns out the other.
When randomness is heavy
Imagine a game where each match is decided by a random roll that the better player can influence only slightly. Suppose the stronger player wins 55% of the time. That is a real, repeatable edge, yet 45% of matches go the other way. Most individual results will not tell you which player is better.
To see how slowly the edge reveals itself, consider a best-of-three between players with a 55% chance per game for the stronger one. The stronger player wins the series if they win two of three games. The chance is the probability of winning the first two (0.55 x 0.55 = 0.3025), plus the probability of exactly one loss among the first two followed by a win. That second case is 2 x 0.55 x 0.45 x 0.55 = 0.27225. The total is about 0.5748, or roughly 57%. Series length matters here, and a short series moves the odds only a little from the single-game figure. The weaker player still wins in about 43 series out of 100.
So the better player has an advantage that is real but small, and anyone watching a handful of series would see an almost even split. This is the central pattern at the heavy-randomness edge: skill exists, but the sample you would need to confirm it is large.
Why a small edge can still matter
A small edge is easy to dismiss, but it is not nothing. Over a very large number of games, a 55% player will come out ahead far more often than not. Ranking systems such as elo ratings rely on this: they assume that a consistent tilt, however modest, will show up in the long run. The edge is simply a different kind of thing from a guarantee. It describes a tendency across many repetitions, not a promise about the next game.
This also explains why games with heavy randomness can still have meaningful skill rankings. It just takes many games to establish them, and the rankings move more slowly and with more uncertainty.
When players are closely matched
The opposite boundary is not about the game being random but about the players being near-equal. If two players are almost identical in ability, then even a game with little built-in randomness produces results that look random, because there is little real difference to express. Small things then decide contests: a single misread, a lapse of focus, a better night’s sleep.
This is why top-level competition can feel more luck-driven than the middle of the ladder. The cause is not that the game got more random. It is that the spread of skill got narrower, so the same small noise matters more in proportion.
Where the design pushes the boundary
Designers move games toward or away from these edges deliberately. A mechanic with a wide spread of outcomes, such as large damage variance on critical hits, pulls play toward the chance boundary, while tight, deterministic interactions pull it away. Neither is wrong. Heavy randomness keeps results open and gives weaker players hope, while low randomness rewards consistency and tends to make results more predictable.
Some systems also soften the extremes. Pity timers in item systems, for example, trim the worst-case tail so that no one is left unlucky for too long, which narrows the spread of outcomes without removing the randomness.
What to be careful about at the edges
A few mistakes are common when people reason about these boundaries.
- Reading a small sample as proof. Near the chance edge, a short run of wins is weak evidence either way. The sample size needed grows as the edge shrinks.
- Assuming no edge means no skill. If the results look even, the likely explanation may be closely matched players, not that skill is absent from the game.
- Assuming a big edge must be visible. A modest advantage can sit unnoticed in a stretch of results, especially over a few dozen games.
- Treating the edge as permanent. Patches, new strategies and changes in the player pool all alter how large the tilt is.
A practical way to think about it
Instead of asking whether a game is skill or luck, it can help to ask how large the skill tilt is per match, and how many matches are needed to see it. A tilt of 55% needs many repetitions to separate from an even split, while a tilt of 80% shows up in a handful of games. Neither number is fixed, and the real figures for any specific game and pairing are usually uncertain.
The boundaries are where this question matters most. When randomness is heavy or players are close, the honest summary of a short result is often just that it was consistent with several explanations, and that more games would be needed to say more.