Comparing Teams Across Eras: A Field Guide
Draft
You cannot put the 2025-26 Knicks and the 1986-87 Lakers on the same floor, so “best ever” is never a fact; it is an argument about method. Comparing teams across decades comes down to two choices, and the ranking can flip depending on which you make. Here is the toolkit, with the 2025-26 Knicks (the biggest raw playoff margin in 43 years) as the worked example.
Two questions you have to answer
- How do you rate the opponents? You cannot credit a team for a tough schedule until you have graded the teams it played.
- How do you put the eras on one scale? A 12-point margin means one thing in a high-scoring, top-heavy league and another in a low-scoring, bunched-up one.
The first is about who they beat; the second is about the world they did it in. They are independent, and a serious comparison settles both.
Axis 1: how you rate opponents
SRS (Simple Rating System: a team’s average scoring margin, adjusted for who it played) is the baseline rating here; Bradley-Terry rates teams from wins and losses alone, the way a chess rating ignores the score of each game; the other two each stress a different assumption.
| Method | Uses | Stresses | Knicks |
|---|---|---|---|
| SRS | margins + schedule | the baseline | 1st |
| Bradley-Terry | wins only | ignores blowouts entirely | 1st |
| Capped-margin SRS | margins clipped at ±15 | no single rout counts double | 1st |
| Elo | margins, recent games weighted more | late-season form | 3rd |
Three of the four make the Knicks #1. Only Elo disagrees, because it credits their opponents for how well those teams were playing late. The lesson: wins-only and blowout-capped ratings are how you rule out “their margins were just padding”; recency weighting is how you ask “but were their opponents tougher than their season record made them look?”
Axis 2: how you put the eras on one scale
| Adjustment | Neutralizes | Knicks |
|---|---|---|
| Raw margin | nothing | 1st |
| Opponent-adjusted | nothing yet; the Axis 1 starting point for the rows below | 1st |
| Scoring-share | total scoring (pace + shooting) | 1st |
| Per-100 possessions | pace only | 1st |
| Spread-standardized | how spread-out the league was | 5th |
Each era adjustment here starts from the schedule-adjusted margin, not the raw one (the same basis the Stats Explainer uses, §22). Start from the raw margin instead and only one row changes: scoring-share drops the Knicks to 3rd.
Two traps live here. Scoring-share scales margins by points per game, which over-corrects: it treats a league’s better shooting as if it were inflation. Per-100 possessions divides by possessions instead, the cleaner fix, because it strips out pace without punishing efficiency. But both only touch the level of scoring. Neither touches the spread of team quality, which has nearly doubled since 1984 (the spread of team ratings rose from 3.1 to 5.9). Measure a team by how far it stood above an average opponent in units of the league’s spread, and a huge margin in today’s top-heavy league counts for less than the same margin in a bunched one.
The punchline: absolute vs. relative
Every absolute measure (biggest margin, schedule-adjusted margin, per-possession margin) makes the 2025-26 Knicks the most dominant champion of the last 43 years. Every relative measure makes them a top-five run: 5th once graded against the league’s spread (the 16–17 Warriors lead at +2.48 spread-widths above average, the Knicks at +1.90), and 3rd once opponents get credit for recent form. Both are honest. They answer different questions:
- “Who had the biggest edge in points?” is an absolute question. The answer is the Knicks.
- “Who was most exceptional for their own time?” is a relative one. The answer is top five, not first.
A rule of thumb
- To rule out a padded margin, switch the opponent rating to wins-only or blowout-capped. If #1 survives, the routs were not the story.
- To handle a high-scoring era, divide by possessions, not points: pace is the artifact, shooting is real basketball.
- To handle a top-heavy era, standardize by the league’s spread. It is the least familiar of the three adjustments, and the one that moves recent teams the most.
- When a claim says “best ever,” ask which axis it used. If it cannot say, it has not compared eras; it has just picked the flattering number.
Built from the 2025-26 Knicks study: full method detail in the Stats Explainer (§5, §9, §15-§22 and “Comparing teams across eras”), full numbers in the Full Numerical Results companion.