1. FC Köln vs SV Werder Bremen

The only two meetings we have point in opposite directions. Köln won 3-1 here on 12 April 2026, a goal up at the interval and three clear by the end. The reverse fixture, at Bremen on 29 November 2025, finished 1-1 with the hosts ahead at half-time.

Four goals and then two. One of the two cleared three and a half, which is the line this preview is interested in, and the other did not come close. Two matches is not a pattern and will not be presented as one, but it is worth noting that neither game was tight defensively in the way the two clubs' league positions might suggest: both saw the away side score, and both were level or better for Bremen at some stage.

The more useful reading is that nothing in the direct record argues against a high-scoring afternoon, and one of the two supports it outright.

The 2025/26 table separated these two by goal difference alone. Köln finished 14th on 32 points from seven wins, eleven draws and sixteen defeats, scoring 49 and conceding 63 for a difference of −14. Bremen finished 15th on the same 32 points, with eight wins, eight draws and eighteen defeats, 37 scored and 60 conceded for −23. Neither was ever close to comfortable, and neither was ever genuinely in the relegation places at the end either. The interesting split is in the attacking numbers: Köln were one of the more productive sides in the lower half at 49 goals, while Bremen's 37 was among the thinnest returns in the division.

This season has produced exactly two matches for each club, and two matches is not form. We will treat it as a sample of two and nothing more. Köln beat Hoffenheim 3-2 at home on 29 August after trailing 0-1 at the interval, then went to Stuttgart on 4 September and lost 4-1, having already been a goal down at half-time. Bremen were beaten 4-1 at Freiburg on 30 August, 2-0 down by the break, and then beat RB Leipzig 3-1 at home on 5 September, leading 1-0 at half-time. Three points each, Köln on −2 goal difference and Bremen on −1.

The first factor is the defensive arithmetic, and it is the most reliable thing on the page. Between them these clubs conceded 123 goals across last season — 63 and 60 — which puts both among the leakiest in the division over a full thirty-four rounds. That is a large sample and it points one way.

The second is what the small sample has done to it. Every one of the four matches these two have played this season has produced at least four goals: 3-2, 4-1, 4-1 and 3-1, nineteen goals in four fixtures. Four matches prove nothing on their own, but they have not contradicted the season-long record either, and when a tiny sample agrees with a large one it is at least not a warning.

Third, the home and away split has been extreme so far, and in mirror image. Köln have won at home and been beaten heavily away; Bremen have been beaten heavily away and won at home. We do not have last season's home and away splits to test that against and will not guess at them, but a goal difference of −23 across a season is not usually spread evenly, and both clubs have now conceded four in a single away match this term.

Fourth, what the fixture means. This is mid-September and nothing is decided, but both clubs know from last season that thirty-two points is close to the line. Meetings between the sides who finished 14th and 15th are the fixtures that decide who spends spring worrying. Neither can approach it cautiously, which tends to add goals rather than remove them.

Fifth, the caveat that applies to every preview written this early. A summer of squad turnover sits between last season's numbers and this match, and we have no way to measure it from the table. The defensive records we are leaning on belong to the clubs, not necessarily to the exact groups of players that will take the field.

We take Over 3.5 Goals at 2.35 with medium confidence. Two bad defences, one weak attack and one reasonable one, and a nineteen-goal start to the season across four matches — the total is where the evidence sits, not the result. Köln at home are worth something, Bremen have already beaten a side that finished third, and picking between them at this stage would be a coin toss dressed up as analysis. The goals market lets us bet the thing we actually have evidence for.

The obvious objection is that 3.5 is a demanding line and the four-match sample carrying it is tiny. The safer version of the identical argument is the 2.5 line, which last season's averages support comfortably on their own: Köln's matches produced 3.29 goals per game across the whole campaign and Bremen's 2.85. We have gone a step further because the price compensates, and because our own settled record says the goals market has been our most reliable and that prices above 2.10 have been the only band we have made money in. Neither of those is a reason on its own, and we are not pretending otherwise.

The thing that beats this pick is Bremen's attack. Thirty-seven goals in thirty-four matches is a side that can go quiet for ninety minutes, and if they turn up to protect a point at a ground where they expect little, a 1-0 or a 1-1 kills the bet by the hour mark. That is the honest risk, and it is why this is medium rather than high.