Scouting scores¶
Research preview. Rank players at the same natural position by expected future league match rating. The full score uses visible attributes, current ability, and up to 5 hidden attributes, including Consistency. The visible-only score uses the attributes shown on the player profile.
Build a shortlist¶
In the web app, open
All players, thenScouting, and choose a position.Set maximum transfer value and weekly wage filters. These are the game’s valuation and the player’s current wage.
Sort by
FullorVisible only. UnderDownload CSV, chooseThis list as shownto keep the shortlist, scores, and scouting position together.
Export the scores¶
fmsave export career.fm players --all --scouting-position CB -o scouting.csv
The export adds scouting_position, scouting_score_full, and scouting_score_visible.
Use --columns to choose which to keep. Scores are blank for players outside the selected
position or with missing inputs.
import fmsave
with fmsave.open("career.fm") as save:
scouting = fmsave.scout_players(save.players(), "CB")
scouting.write_csv("scouting.csv")
Use to_pandas() or to_polars() on the same table. To score a player directly, call
fmsave.scouting_score(player, "CB", model="full"). The 8 options are GK, CB, FB/WB,
DM, CM, AM, WIDE, and ST. For a player with several natural positions, score
each position separately.
How the scores were tested¶
Each position has a weighted sum of attributes. The weights were fit to future league ratings and checked by holding out each of 5 FM26 careers in turn. Across 2 seasons of 1 FM27 career, the full score reached a rank correlation of 0.570 within league and position, compared with 0.504 for current ability. Center backs showed the largest gain.