Strength of schedule explained
A team's record means little without knowing who it played. How strength of schedule is calculated, why it matters most in college sports, and how EdgeIQ uses it without leaking future results.
By Patrick C · Founder & Editor, EdgeIQ
In this article
Two teams with identical records can be very different. One may have built its record against the league's best; the other against its weakest. Strength of schedule, often shortened to SOS, is the attempt to measure that difference so a record can be read in context.
The basic idea
SOS summarizes how good a team's opponents were. The simplest version averages opponents' win percentages. More refined versions weigh opponents' own opponents, use point margins instead of records, and account for where games were played.
| Method | How it works | Weakness |
|---|---|---|
| Opponents' win % | Average record of teams faced | Opponents' records are themselves inflated or deflated by their schedules |
| Two-level (opponents + opponents' opponents) | Blends both layers, as the NCAA's old RPI did | Still ignores margin |
| Rating-based | Average rating (e.g. Elo) of opponents | Only as good as the ratings |
| Margin-based | Opponents' average point differential | Blowouts can distort it |
Why it matters most in college sports
Professional leagues are designed for balance: salary caps, drafts and structured schedules keep teams relatively close. College football and basketball have hundreds of programs with enormous resource gaps and schedules that differ wildly. A 10–2 record in one conference can mean something completely different from 10–2 in another. That is why college rankings and tournament selection lean so heavily on schedule strength.
How EdgeIQ measures it
EdgeIQ's engine includes a strength-of-schedule feature built from the average margins of each team's past opponents, computed only from games that were completed before the game being predicted. The Elo-style ratings add a second, implicit layer: beating a highly rated opponent raises a rating more than beating a weak one.
Avoiding a subtle leak
It is tempting to compute SOS using opponents' full-season records. But when predicting a game in October, the full season has not happened yet. Using end-of-season opponent records would quietly give the model information from the future and make backtests look better than real life. EdgeIQ rebuilds SOS as of each game date so every number reflects only what was known at the time.
Strength of schedule is only useful if it is computed as of the prediction date. End-of-season SOS in a backtest is a form of data leakage.
A worked example
Team A is 6–2 and its opponents have averaged a −3 point margin against everyone else. Team B is 5–3 and its opponents have averaged +4. On record alone, A looks better. Adjusted for schedule, B has been winning against considerably stronger opposition, and a model that accounts for SOS may rate B as the stronger team.
Limitations
- Early in a season, opponents' numbers are themselves based on a few games, so SOS is rough.
- Schedules with few shared opponents — common in college sports — make comparisons between conferences uncertain.
- SOS describes the past schedule; the difficulty of a team's remaining schedule is a separate question.
How to use SOS as a reader
When a team's record looks surprising, look at who it has played. On EdgeIQ game pages the factor breakdown shows when schedule strength is pushing a projection one way or the other, and team pages show recent opponents so you can judge for yourself.
Sources and further reading
About the author
Patrick C founded EdgeIQ and edits its Learning Center. He oversees how EdgeIQ collects sports data, how its prediction models are evaluated and how results are explained to readers.
EdgeIQ is a sports analytics platform. Projections are statistical estimates based on historical data, carry uncertainty and are never guarantees. EdgeIQ takes no wagers and is not a sportsbook. Read the methodology and analytics disclaimer.