Individual points percentage is a linemate stat
Individual points percentage — IPP, the share of goals scored while a player is on the ice that he collects a point on — has become a favorite tool for explaining why a player’s point totals rose or fell. A high IPP is read as deep involvement, a sign the offense runs through him; a low one is read as a passenger riding his linemates’ production. Both readings assume IPP is a stable trait of the player, a measure of how central he is to the attack. It is not. IPP is one of the most linemate-dependent and randomness-prone numbers in hockey, a ratio built on a small denominator of on-ice goals that swings wildly year to year and regresses hard toward the middle. It tells you as much about who a player skates with and how the bounces fell as it does about the player himself.
What IPP actually records
IPP is a simple ratio: the points a player records divided by the total goals his team scored while he was on the ice, expressed as a percentage. An IPP of seventy means the player earned a point on seventy percent of the goals scored during his shifts. It is meant to capture how involved a player is in the offense that happens around him.
But the denominator is small — a player is on the ice for a limited number of goals in a season — and the numerator depends on whether the scorer or the passers on any given goal happened to be him. A ratio of two modest counts, one of them heavily influenced by chance, is a fragile thing, and IPP inherits all of that fragility.
Linemates set the ceiling and the floor
Who a player skates with drives his IPP more than almost anything he does. A forward on a line with two elite finishers will see plenty of goals scored during his shifts that he had no hand in — a give-and-go between his linemates, a shot he was nowhere near — and every one of those drags his IPP down without reflecting his ability. A player who is the clear focal point of a line takes a bigger share by role.
This means IPP partly measures a player’s position in the pecking order of his line, not his quality. Move a good playmaker onto a line where he is the third option and his IPP falls; make him the trigger man and it rises. The number reflects the distribution of touches within a unit, which is a function of linemates and deployment as much as of the player.
Deployment shapes the denominator
How a player is used changes the goals he is on the ice for and therefore his IPP. A power-play specialist collects points on a high share of the goals during his heavily skilled minutes, inflating his IPP, while a defensive forward buried in his own zone is on the ice for goals in transition he had little part in. The role assigned by the coach seeds the ratio before the player does anything.
Even strength of competition matters: a player deployed against weak opposition in offensive situations will see more goals go his way in ways that pad the numerator, while one sent out to defend leads is on the ice for empty-net goals and scrambles that distort the denominator. IPP does not adjust for any of this. It takes the goals a player’s usage produced and asks only what share he touched.
The randomness nobody prices in
Because IPP rests on a small number of on-ice goals, single-season figures are noisy, and the noise is easy to mistake for signal. Whether a player picks up the secondary assist on a goal — a point that hinges on who last touched the puck two passes earlier — is close to random, and secondary assists make up a large share of what moves IPP. A player can gain or lose several points of IPP on nothing but the sequencing of passes on goals he was near.
This is why IPP regresses so hard. A player with an unusually high or low IPP in one season tends to drift back toward the league norm the next, not because his role changed but because the initial figure was partly luck. Treating a one-year IPP as a fixed attribute is exactly the mistake the regression warns against.
Reading involvement without IPP
The thing IPP gestures at — how much a player drives offense — is better captured by numbers tied to what he actually did. Primary points per sixty minutes, which counts goals and first assists while stripping out the luck-laden secondary assist, isolates real involvement far better than a ratio built on total on-ice goals. Individual shot and chance generation shows whether a player is creating offense or standing near it.
On-ice shot-share and expected-goals numbers describe whether the offense actually improved with him out there, independent of who finished. Read together, these tell you what IPP only hints at: whether a player is a genuine engine of his line’s scoring or a beneficiary of it. IPP by itself cannot make that distinction, because it counts touches, not creation.
What to read instead
Use IPP, if at all, as a flag rather than a finding. A wildly high IPP is a hint that a player’s point total may be unsustainable and due for regression; a low one on a productive line hints there may be more scoring to come. In both cases the honest move is to check the underlying rate stats rather than to conclude anything from the ratio alone.
Primary points per sixty, individual chance creation, and on-ice expected goals give a durable picture of involvement that survives from one season to the next. IPP is the volatile surface number; those are the signals underneath it. When they disagree, trust the ones with the larger, less luck-driven samples.
The honest read
Individual points percentage is a real ratio that captures something intuitive — the share of nearby goals a player has a hand in — and at the extremes over a long stretch it carries a little information about a player’s offensive role. A star who consistently touches most of his line’s goals is telling you he is central to it.
But IPP is dominated by linemates, deployment, and the randomness of secondary assists, and it regresses hard enough that a single season’s figure is a weak foundation for any claim. Read it as a linemate-and-luck stat that flags players due to regress, not as a measure of how good a player is — and when you want real involvement, read the primary-point and chance-creation rates it sits on top of.