The pick that wins your pool isn't the pick that wins the game
Every other pick'em tool ranks this week's games by who is likely to win. That is a perfectly good answer to a question nobody in a pool is actually asking. You are not trying to be right. You are trying to finish first — and once there is a field between you and the money, those two things come apart.
Where they come apart
One game left. You are a point behind the leader, and the leader has already locked the home favorite — an 80% shot.
- · Take the favourite and you move together. They stay ahead whether it wins or loses. Your chance of finishing top is 0%.
- · Take the underdog and you win the game 20% of the time — and in exactly those cases you catch them. Your chance of finishing top is 20%.
The correct pick is the one that will probably lose. Nothing about that follows from a better forecast — it follows from asking a different question.
Survivor is one problem, not fifteen
A survivor team can be used once. Taking the 96% favorite in week 3 is also removing them from week 11, where the board is thin and they were the only comfortable ride left. So the real question is not “who wins this week” but which assignment of teams to remaining weeks survives all of them. We solve that exactly, rather than picking the best team each week and hoping.
Chance of surviving every remaining week, computed on the current schedule from our own win probabilities. Both numbers are calculated when this page builds — neither is written into the copy, so neither can quietly stop being true.
It never sees anyone's hidden picks
A well-run pool hides picks until kickoff, and ours does. The strategy engine is built so that it cannot read a pick that has not locked — it holds no database connection at all, and there is nowhere in it to put one. It works from what everyone in your pool can already see: the standings, the games left, the market line, and how each member has picked in weeks that are already revealed. Survivor planning needs even less than that — it is entirely about your own remaining teams.
What we don't claim
We do not beat the closing spread. We have measured it on our own numbers and published the result, including the unflattering parts. Anyone selling you 60%+ winners against a closing number is selling you something. None of the above depends on out-forecasting the market — it depends on optimising a different objective, which is why it works even though the market is efficient. The record is here, losses and all.
The engine runs inside every pool on Gridpex. Start one — it is free, there is no entry fee, and we never hold money.