FMAX Tactic Lab
How tests work
Under what conditions the Tactic Lab’s scores are measured, and how. Worth one read before you read the leaderboard.
- 1The same test for everyone
Everyone tests with the same clubs, the same squads and the same opponents.
- 2Each in their own game
The matches are played by the tester’s own game, and FMAX collects the results.
- 3Many people’s values, pooled
All values measured for the same tactic are pooled. Each tactic shows how many people measured it.
The conditions of the official test
In an official test FMAX has the game play matches for you to measure how strong a tactic is. The conditions are made the same so that results can be compared side by side, whoever measured them and wherever.
- Clubs
- The test uses the two clubs set on the test sheet. The tester cannot choose. In a save where one of them cannot be used because you manage it, a stand-in club set to the same conditions takes its place.
- Squads
- Everyone tests with the same official squads. The squads are swapped only for the test and put back when it ends, so the players of your save stay as they are.
- Opponents
- The tactic plays the game’s own ready-made tactics, home and away in equal numbers. It meets opponents of many styles, so a tactic strong against only one kind struggles to score high.
- Match engine
- Only values measured on the same match engine version go to one leaderboard. When a patch changes the match engine, a new leaderboard starts and the old one is kept as a record you can choose by version. Testers do not pick the version: FMAX reads it from the game.
- Live tweaks
- FMAX’s Live tweaks that affect matches are switched off during a test.
Choose one of three sizes
The more matches, the narrower the margin of error. Several tests of the same tactic are all added together.
Each person measures in their own game
- The matches are played by the game on your computer
- The Tactic Lab’s server plays no matches. The tester’s Football Manager 27 plays them, and FMAX collects the results.
- Every finished test is recorded
- When an official test ends, a summary of the result is recorded in the Tactic Lab by itself. This keeps people from uploading only the runs that went well. The tactic file is sent only when you press ‘Upload to the Tactic Lab’.
- The tactic is fixed before the start
- Once a test has started the tactic cannot be changed, and results played with another tactic cannot be uploaded as this one’s.
How many people’s values are pooled
- One person, one voice
- However many matches one person measures, they count as one person’s share. On a tactic measured by several people, nobody can decide the score alone.
- Ranked from 100 matches
- A tactic goes onto the leaderboard once it has 100 matches or more in total. Until then only its values show under ‘Recently uploaded’.
- Tactics with one tester are marked
- A tactic measured by one person only is listed too, marked ‘One tester’. When someone else measures it the mark goes away and the score becomes more trustworthy.
- Values that disagree are marked
- When the uploader’s value disagrees with other people’s beyond the margin of error, the tactic is marked ‘Not reproduced’ and scored on the other people’s values.
- Outliers move it less
- The values are pooled so that one or two unusually high or low ones do not move the overall score much.
How to read the scores
- Points
- The points expected from one 38-match season with this tactic. Higher means stronger.
- Margin of error
- How far luck can move the result. The true points are almost certainly inside this range. More matches make it narrower.
- Rank
- Tactics are placed 1st, 2nd and so on by what has been measured so far. Tactics whose margins of error overlap may change order with more matches, so beside the rank you also see where the tactic is likely to be, such as ‘est. 2nd–5th’.
- Equal points
- There are no shared ranks. The tactic with the better goal difference per match comes first; if that is equal too, more goals per match, then a higher share of wins, then more matches measured, then the earlier upload.
- All measurements, or official conditions only
- The leaderboard first shows the pool of every measurement received. Choose ‘Official conditions only’ for the pool of results measured with unmodified game files and the official test’s own clubs.
Grouping similar versions
- Automatic grouping
- When one user uploads several slightly tweaked versions of the same tactic, those versions can quickly fill up the leaderboard. Following user feedback, Tactic Lab automatically groups similar versions together, so uploaders do not need to do anything extra.
- What counts as the same tactic
- Versions must come from the same account; tactics uploaded by different users are never grouped, no matter how similar they are. They must also share the same authorship info (whether the uploader made the tactic themselves, uploaded someone else’s tactic, or made a Tweak of someone else’s tactic, along with the creator’s name). All 11 player positions must match both in possession and out of possession, and roles and instructions must be nearly identical, with only a few settings changed. If even one player is in a different position, it counts as a separate tactic. When one version is similar to a second and that second version is similar to a third, all three go into the same group. Tactics that are unpublished, hidden, or closed to downloads are not grouped.
- Top version
- Each group is shown on the leaderboard through a single “top version”. This is the version with the highest points even after accounting for its margin of error. Because a version with only a few matches has a wide margin of error, a lucky early run with high points will not immediately make it the top version. The top version is not fixed: it is chosen anew whenever the leaderboard loads, so it can change as more matches are played and the numbers shift.
- Ranks and formation leaders
- On both “Leaderboard” and “Standard leaderboard”, each group takes up a single row, and both rank numbers and estimated rank ranges (such as 2nd to 5th) count each group as one tactic. The “Best by standard formation” and “Best by formation” tabs, along with first-place formation labels on the leaderboard, also count each group as one, so versions from a single user do not take up multiple spots in the same formation.
- Viewing and comparing versions
- Clicking “similar versions” on a top version’s row expands the other versions right beneath it, showing each version’s rank across all versions alongside its own stats. Clicking “different settings” on an expanded version shows the settings that differ from the top version with both versions’ values side by side. Expanded groups stay in the page address, so sharing that link opens the exact same view for others. Above the table, “Show all versions” gives every version its own row and ranks each one separately, while “Group similar versions” switches back.
- Search, filters, and recent uploads
- When a search or filter matches only a version other than the top version, that version appears on its own row and the rank column notes which ranked tactic it is another version of (for example, another version of the 3rd-place tactic). The “Recently uploaded” tab does not group versions and lists every upload in order.
- Individual stats and tactic pages
- Stats are never combined across versions. Every version keeps its own points, margin of error, match record, matches, number of testers, downloads, comments, tactic page, and web address, and existing links still open as before. Each grouped tactic’s page includes a “Versions of this tactic” table showing the points, rank, and matches for every version, with the top version marked “Top”. Pages for other versions also include a note explaining that they appear grouped under the top version on the leaderboard.
Limits to keep in mind
The Tactic Lab’s scores are values many people each measured and pooled, not records checked match by match.
- It is a rank in the official test
- The order is the order measured with the official test’s clubs and squads. With your team’s players and your opponents it may differ.
- Small differences cannot be told apart
- Good tactics are often only a few points apart. Even with different rank numbers, when the margins of error overlap it is not yet known which is stronger.
- It may differ a little between saves
- The conditions are made the same, but some difference may remain from save to save. That is one reason many people’s values are pooled.
- Made-up values cannot be ruled out entirely
- So each tactic shows how many people measured it, and the pooling moves closer to the true value as other people’s measurements build up.