This forum is archived and no longer active. You can visit us in our Discord Server here!

Topic: Projected Scores

1

Is there a way to see every possible score for a projected score range. I like to limit my player pool down to around 20-25 players each NBA night but it seems like the best actual score for that player pool is never in the projected lineups page even when I expand it out too 100 lineups. Thanks in advance.

Short answer, No. Longer answer Nooooo. Even longer answer:

Hypothetically if you have 25 players, you would have 5 of each position. The total combinations of that many players is 50,000. You could eliminate some of those lineups because they would be over the salary cap, the exact number would depend on which guys were in your pool. Our algorithm is fast because it finds the optimal solution without ranking every single possible outcome. We just aren't setup to output 50,000 lineups currently. Even more so, if you did look at the 200 surrounding the actual optimal score, the other 200 would likely be all duds.

Lets say the projected optimal score for today is 310.2. The next lineup may be 310.1, the next 310.0 and so on. However you will notice as the number falls, there are more combinations that produce the exact same number. So when you get to say 305.6, you may have 10 unique lineups to produce that same number, and the further down you go, the more ways you can combine the players to get the exact same projection.

If you limit your pool to 25 on lineup rewind and get the best actual score, you will see the projected score of that lineup as well. You can tell how far away that particular lineup was from the best possible projected lineup. If the projection is much lower than the best projection score then you would likely find hundreds of lineups that had the same projection total. You would not be able to determine before the day which one of those hundreds of lineups would be the winner, let alone if you even guessed the correct range to start with.

Hope this helps clear it up a little.