15,000 matches to answer one question — do tactics matter?

Every management game says its tactics matter. Few show you how much. In September 2026, while checking the claims on our own website, we decided to measure ours properly.

How we tested

The idea is simple, the discipline is not. We took one team and played blocks of 1,000 simulated matches against the same opposition, changing one setting at a time and comparing the win rate with a reference block. In total, 15,000 matches. Nothing was saved to any real world; the simulations existed only to be counted.

Before looking at any result, we wrote down what we expected to see and how big a difference had to be before we would believe it.

The noise floor

The first thing to measure is how much results move when nothing changes at all. We ran the identical, neutral setup several times. The win rates came out at 37.6%, 38.1% and 39.8% — a spread of about 2.2 percentage points from pure chance.

That number is the most important one in this post. Any effect smaller than about two points cannot be told apart from luck, however good it looks in a single run.

What we found

• Playing style was by far the biggest lever. A direct, vertical style beat a patient possession side by 15 to 19 percentage points — far beyond the noise.
• A high defensive line helped, by around six points: a real effect, but close to the edge of what we could confirm.
• Maximum pressing, faster tempo and a deep line all stayed inside or near the noise floor. Pressing in particular lost most of its benefit in squads with low stamina — which is realistic, but meant it rarely paid off in lower divisions.
• Three effects described on our homepage at the time — about the defensive line against fast forwards, the cost of pressing, and tempo — did not show up at all. We removed them from the site.

Then the twist

A style that wins by fifteen points looks like an exploit. So we tested it again, this time against a realistic mix of opponents rather than possession sides only. The vertical style still beat four of the five styles we tested, by 13 to 29 points — and then lost to high pressing by 39 points.

Weighted by how often each style actually appears in the game, vertical play still came out ahead, by eight or nine points overall. But it is not an upgrade; it is rock-paper-scissors. The right style depends on who you are playing.

What changed afterwards

The experiment had a consequence we did not expect. For the style matchups to mean anything, AI clubs need to actually play different styles — and at that point, most of them were not choosing tactics at all. That is the subject of the next entry, AI clubs finally choosing their tactics.

What it means when you play

• Look at the opponent's style before each match, and pick yours with it in mind. That is where most of the tactical edge is.
• Treat the four dials as fine-tuning. They are worth a few percent, and only in specific matchups; the tactical dials article lists exactly where.
• Judge changes over several matches, not one. With a noise floor of two points, a single result tells you very little.

Next in the diary: AI clubs finally choose their tactics.