Deployment jumps

Breakout Comps: he just got a bigger role. What happened to players like him?

The latest PP1 promotions, lineup moves and ice-time jumps, each matched with the historical players who made the most similar jump from the most similar starting point, and what they produced over the next 20 games.

Latest jump Nov 15, 2026 · comp pool from 2018-19, 2019-20, 2020-21, 2021-22, 2022-23, 2023-24, 2024-25, 2025-26 (older seasons are still loading, so the pool grows over time).

Latest jumps

Show

How the comps are built

  1. A jump is a promotion from our deployment tracker (PP1, a line move up, or an ice-time spike) or a sustained change we detect ourselves: TOI up 2:30 or more across 3 straight games, or power-play share up 20 points.
  2. The pre-jump profile covers the 10 games before: TOI, power-play share, expected goals per 60, shot diet (xG per unblocked attempt), points per game and age.
  3. Comps are the 8 nearest historical jumps at the same position group on those features plus the size of the jump, one per player, among jumps with a full 20-game outcome.
  4. We report what the comps did over the next 20 games (points per game and the change from before), and whether they kept at least half of the jump in games 11 to 20.

Expected goals use our trained xG model where games have been processed, and shots × league shooting % elsewhere. Ages come from NHL birth dates.

FAQ

What counts as a breakout jump?

A promotion to the first power-play unit, a move up the forward lines or defense pairs, a one-game ice-time spike our deployment tracker flags, or a sustained jump of 2:30+ of ice time over three games or 20+ points of power-play share.

How are comps chosen?

By nearest neighbours on a standardised feature vector: age, ice time, power-play share, expected goals per 60, shot diet and points per game before the jump, plus the size of the jump in ice time and power-play share. Comps must play the same position group and have a complete 20-game outcome.

Which seasons are in the comp pool?

Every season we have loaded (the page lists them), weighted toward recent seasons. A jump from last season counts 80% as much as one from this season, two seasons ago 60%, then half again for each older season. Older jumps need to be a closer match to make the list, and the median outcomes lean on the recent comps, because roles and scoring levels drift.

Does keeping the role matter?

Yes. Most one-game promotions do not stick. The kept-role rate tells you how often similar jumps lasted, which is usually the difference between a pickup and a trap.