[shinsengumi]

Worst case, if Gauss doesn't show up, I'll do it - as long as everybody promises not to complain if when I'm not punctual about posting results or that it does not look as slick as Gauss's python generated tables.  Mine will look like I used Karel The Robot to generate the tables.

Where have you gone Gauss?  Our message board turns its lonely eyes to you.

Can somebody confirm that PLF's recollection of the scoring system is correct?  The system worked pretty well so let's stick to it.  Whatever it was.

[quote author=PrinceLightfoot link=topic=10428.msg95138#msg95138 date=1408939903]
If I remember correctly, we ended up following this formula: [Stanford predicted score - Stanford actual score] + [Opponent predicted score - Opponent actual score] = margin of error. The total number of points awarded is based on the number of players entering in a prediction for that week...i.e. 20 predictions, the person with the smallest margin of error gets 20 points, 2nd place gets 19 points, etc. There was a 10(?) point bonus if your score prediction was perfect and a 5(?) point penalty for picking against Stanford with the result being a Stanford win. Your weekly score would be added to the season-to-date totals.

[/quote]

Publius, I am happy to help you out in this year's effort. I can take a first stab at setting up the scoring model. Let me know if that works.

I think PLF's recollection of the scoring is pretty much spot on. My one recommendation would be to get rid of the "Stanford Homer" rule to level the playing field for our outside guests (Miller, Gemmel, Vegas Line, etc.). Instead, we can award 5 additional points for predicting the correct winner. As it stands, someone might pick the wrong winner in a close game, but end up getting more points compared to someone else who picked the right winner.

[shinsengumi]

Worst case, if Gauss doesn't show up, I'll do it - as long as everybody promises not to complain if when I'm not punctual about posting results or that it does not look as slick as Gauss's python generated tables.  Mine will look like I used Karel The Robot to generate the tables.

Where have you gone Gauss?  Our message board turns its lonely eyes to you.

Can somebody confirm that PLF's recollection of the scoring system is correct?  The system worked pretty well so let's stick to it.  Whatever it was.

[quote author=PrinceLightfoot link=topic=10428.msg95138#msg95138 date=1408939903]
If I remember correctly, we ended up following this formula: [Stanford predicted score - Stanford actual score] + [Opponent predicted score - Opponent actual score] = margin of error. The total number of points awarded is based on the number of players entering in a prediction for that week...i.e. 20 predictions, the person with the smallest margin of error gets 20 points, 2nd place gets 19 points, etc. There was a 10(?) point bonus if your score prediction was perfect and a 5(?) point penalty for picking against Stanford with the result being a Stanford win. Your weekly score would be added to the season-to-date totals.

[/quote]

Publius, I am happy to help you out in this year's effort. I can take a first stab at setting up the scoring model. Let me know if that works.

I think PLF's recollection of the scoring is pretty much spot on. My one recommendation would be to get rid of the "Stanford Homer" rule to level the playing field for our outside guests (Miller, Gemmel, Vegas Line, etc.). Instead, we can award 5 additional points for predicting the correct winner. As it stands, someone might pick the wrong winner in a close game, but end up getting more points compared to someone else who picked the right winner.