How to build better tournament lineups
The one mental shift that changes everything: you are not trying to be right, you are trying to be right in an unusual way.
A cash game and a large-field tournament are different games that happen to use the same roster rules. In a cash game you need to beat about half the field, so the best lineup is the one most likely to score well. In a tournament with ten thousand entries you need to beat all of them, and the lineup most likely to score well is also the lineup thousands of other people built.
That is the whole problem. Being right is not enough if everyone else was right in the same way, because the prize gets split. What pays is being right in a way the field was not.
Four things that follow from that
- Build a portfolio, not a lineup. One tournament lineup is a lottery ticket. A set of lineups is a position — it can cover several ways the day could go, and the ones that miss cost you the entry fee while the one that hits pays for all of them.
- Correlate. Uncorrelated lineups score close to their average. Correlated ones have fat tails in both directions, and in a tournament only the top tail is being paid. See How stacking works.
- Use ceiling, not just projection. Winning a big field takes an outcome near the top of a player’s range, not his typical game. A player projected for 12 who can score 30 is worth more in this format than one projected for 14 who tops out at 18.
- Do not fade everything popular. The overcorrection is as expensive as the mistake. Winning large-field lineups have historically carried plenty of popular players — they are popular because they are good. The edge comes from being different in a few well-chosen places, not everywhere at once.
Doing it here
- Start with Just Pick For Me or Balanced at a real portfolio size — 20 lineups at minimum, 150 if you are entering a large field. Both weight ceiling and correlation, not projection alone.
- Open the Stacks tab and check the mixture is one you can defend. If you want more of a particular shape, add a stack rule rather than hand-building lineups.
- Open the Exposure tab and find the two or three players you genuinely disagree about. Turn each disagreement into one rule. Two good rules beat twenty guesses.
- Build again and compare the two with Saved builds so you can see exactly what your rules did.
None of this makes a portfolio a winner. Football is played by people and most weeks most entries lose. What good construction does is stop you losing for reasons that had nothing to do with the games — duplicated lineups, one bad player in everything, no exposure to the outcomes you actually believed in.