Comparing two teams through a common opponent rarely works

The shortcut and its appeal

When two teams are about to meet and have no recent history, the obvious move is to compare how each performed against an opponent they both faced. The comparison feels rigorous because it holds something constant, and holding something constant is the basic move of any fair test. The problem is that almost nothing else was constant, and the differences are large enough to dominate whatever the comparison appears to show.

Because the reasoning looks structured, conclusions drawn this way are stated with far more confidence than the method can support. The shortcut is not useless, but it requires so many caveats that it usually collapses into a much weaker statement than the one being made.

The common opponent was not the same twice

A team that plays two fixtures weeks apart is not identical on both occasions, since availability, structure and physical condition all change. It may have been rotated heavily for one and fully loaded for the other, which is enough on its own to invalidate the comparison. Its own approach may have differed too, since teams adjust their setup to each opponent and the two setups may have been very different.

Where the two fixtures occurred at different points in the season, the intervening period may have included substantial tactical change. Treating the common opponent as a fixed measuring instrument is therefore the first and largest error in the whole procedure.

Circumstances differed on both sides

One of the two performances may have occurred away from home in a congested week and the other at home after a long break. Those differences typically matter more than the modest gap in performance that the comparison is attempting to detect. Weather, surface and officiating all vary as well, and each can push a single match some distance from what either side would normally produce.

A comparison that ignores these factors is attributing to team quality a difference that context can comfortably explain on its own. Correcting for them is possible but requires so much adjustment that the original observation contributes very little to the final estimate.

Matchups break the chain

Even with identical circumstances, transitivity fails because performance against an opponent depends on the interaction between the two styles involved. A team may handle the common opponent easily because that opponent's approach suits them, while the other team struggled for exactly the opposite reason. Neither of those facts says anything about how the two will handle each other, since that is a third and entirely separate interaction.

This is the deepest reason the shortcut fails, and it persists no matter how carefully the contextual differences are controlled. Style relationships simply are not transitive, and no amount of adjustment can make them behave as though they were.

What to do instead

The more useful move is to describe each team's structure directly and then reason about how those two structures would interact. That approach requires more work and yields less quotable conclusions, but it does not depend on a chain of inference that is known to break. Common opponents remain useful as sources of observation about how each team behaves under particular conditions, which is a different use entirely.

Watching how a side handled a high press tells you about that capability regardless of who applied the press or what the outcome was. Used for observation rather than for ranking, the same fixtures become genuinely informative without any reliance on transitivity.