Köln2-2Hertha

This preseason friendly at RheinEnergieStadion gives Köln and Hertha a chance to tune up before competitive action resumes. The two sides have met only twice before on record, and both of those meetings were decided by a single goal.

The most important numbers here come from the Poisson model, which assigns Köln a 48.1% win probability, a 20% draw probability, and a 31.9% Hertha win probability. The projected scoreline is 2-1. Rest days are a clear differentiator: Köln are eight days removed from their last match, while Hertha last played on May 17 and open their season here.

Köln's home record over the last 20 matches at RheinEnergieStadion shows 7 wins, 5 draws and 8 losses, a win rate of 35%. They have scored an average of 1.9 goals per game and conceded 1.6.

Their most recent five matches across all competitions have been higher-scoring, with an average of 2.6 goals scored and 2.4 conceded per game. The form sequence reads W-L-L-D-L, but the sample is small and includes a match just eight days before this fixture.

Hertha's away record over the last 20 matches is stronger on its face: 10 wins, 4 draws and 6 losses, a 50% win rate. They average 1.75 goals scored and 1.5 conceded on the road.

Recent form tells a different story. Over their last five matches, Hertha have one win, one draw and three defeats, scoring an average of 0.8 goals and conceding 2.0. They have not played since May 17, a 75-day gap that makes this their season opener.

Only two prior meetings between these clubs are on record. Köln won 1-0 at Hertha in November 2024, and Hertha won 1-0 at Köln in April 2025. Both matches finished with a single goal, and in both cases the away side came out on top.

The combined goal total across the two meetings is two, an average of one per game. With such a small sample, the head-to-head offers little predictive value, though the low-scoring pattern stands in contrast to the model's projected 2-1 scoreline.

Köln's home win rate of 35% over the last 20 matches at RheinEnergieStadion is lower than Hertha's away win rate of 50% over the same stretch. Hertha also concede slightly fewer goals on the road (1.5 per game) than Köln concede at home (1.6 per game).

Despite these splits favoring Hertha at first glance, the model still gives Köln the higher win probability. The model's medium confidence leaves room for error, and the venue data alone does not decisively favor either side.

The clearest factor flagged by the data is rest. Köln last played on July 23, giving them eight days between matches. Hertha last played on May 17, a 75-day gap that makes this their first match of the new season. The extended layoff introduces additional uncertainty into any projection for Hertha.

The head-to-head sample is also small — only two meetings on record — so past results cannot be weighted heavily. Finally, the model's confidence is medium, meaning the probability estimates are not precise. Any of these factors could push the actual result away from the projected 2-1 scoreline.

The Poisson model gives Köln a 48.1% chance of winning, a 20% chance of drawing, and a 31.9% chance of Hertha winning. The projected scoreline is 2-1, based on expected goals of 2.28 for Köln and 1.84 for Hertha.

Among exact scorelines, 2-1 is the most likely at 7.8%, followed by 2-2 at 7.2% and 1-1 at 6.8%. It is important to note that the most likely outcome, as the model tallies it, is a Köln victory — the aggregated win probability exceeds the probability of any single scoreline. The model rates its own confidence as medium, so these numbers are best viewed as a central projection rather than a precise forecast.

The data points to a home win as the most probable outcome, with a 48.1% win probability and a projected 2-1 scoreline. The combination of Köln's more recent match activity and Hertha's 75-day layoff underpins the model's lean toward the home side.

However, the medium confidence and Hertha's superior away record over the last 20 matches mean the result is far from settled. An away win is not unlikely at 31.9%, and the draw is at 20%. The model's projected scoreline implies more goals than the two prior meetings produced, so a repeat of the low-scoring head-to-head pattern would be the clearest way the projection misses.