A simple question: We all know that when setting up a box pool, you don’t draw the numbers until the grid is filled in. However, when I brought the chart I made for our office pool into work this morning, one person cried “foul” because I had already labeled one side as the Giants and one side as the Patriots. He claimed that the “real” way to do it was to randomize that as well.
In all my years of running/participating in box pools, I’ve never once heard of anyone doing this. While I recognize that it would add another layer of randomization, that layer seems to be completely unnecessary to me.
So what say you – label the teams before or after the chart is filled out?
I say it’s 100% irrelevant either way. With no numbers, it doesn’t matter what team is on what side. But all the one’s I’ve done have had the teams already on it.
Well, the only thing I can think of is if someone is positive that one of the teams is going to get shutout, and wants to make a 1/10 shot at locking down all the zeros for one particular team. (Or if they think one team will get a FG early, and then nothing else, etc.)
It doesn’t matter at all whether the sides are marked with the team names or not. Since you don’t know which numbers are going to go into the columns and rows, there is no strategic advantage in picking any one of the 100 squares–because, in a box pool (or “football squares” as I’ve often heard it), it’s the numbers, not the identities of the teams, that makes the difference. Analysis of NFL final scores will show you that some numbers are more or less likely to appear at a quarter’s or game’s end, but that no individual team is more or less likely to get one of those numbers than another.
If you really want to people to cry foul in football squares, make sure the administrator of the pool gets all the 0’s, 3’s, and 7’s. Outrage will certainly ensue.
Tell him you’ll flip a coin. Heads, it stays the way it is. Tails, you’ll rotate it 90 degrees, which would have **exactly the same effect **as randomizing the team names.
I used to print out a huge super bowl grid (on a blueprint printer). It is amazing how many dumb questions are asked. Same people, same questions.
But no one ever complained that I labeled one team on top and another team on the side.
One guy was convinced that if one team lost, he could not win.
Some advice, If you are dealing with a lot of people who have been drinking, don’t make a 5x5 square and try to explain that you get four squares for the price of ONE.
Yeah, somehow early in my career I became the office pool guy. I’ve gotten that complaint a few times. Luckily there was usually someone else in my office ready to laugh at the person who complained.
In a similar vein one of my jobs is running medical trials. I discovered that if you need to have 5 vials of medicine and 5 vials of placebo you can’t just label the vials 1 through 10 and assign medicine and placebo to them randomly without someone complaining. Instead you need to label the 10 vials with random 5 to 10 digit numbers and then assign medicine and placebo to them randomly. Apparently if they have sequential numbers it’s less random. >_<
That wouldn’t matter anyway, even if the guy was right about his prediction. Suppose I had divined that the Giants would get shut out. I could either grab a column of Giants and hope that I make the 1/10 shot of that column being ‘0’, as you suggest - OR- I could also grab a horizontal row of Patriots, knowing that no matter what, I get a single square of Giants=0, no matter where the Giant/0 column falls. In that case, I have a 1/10 shot of getting the matching Patriot square.
So your way, I definitely get the Pats score and a 1/10 shot of getting the Giant=0, or the other way, I definitely get the Giant=0 and have a 1/10 shot of getting the Pats score.
In case anyone is interested here is a link to a study of most likely scores. Not surprisingly 0,3,4,and 7 combos dominate while those with 2 tend to be bad.
Ah-yup. I ran a box pool for the Giants/Niners game – the lucky bastard who got NYG-7/SF-0 ended up winning the 1st and 2nd quarters as well as the final score.
Payout by quarters introduces a correlation between quarter-wins in adjacent rows and columns, because modulo-0 is a common (>10%) quarter-score for a team. If a team is low-scoring, they will have a greater degree of inter-row (or inter-column) correlation because they will quarter-score 0 more often than a high-scoring team.
So if you want outcome correlation (because you have a multiple-wins-or-nothing mindset), and you think that one team will score more than another, then you care which team is on which axis. You want to string out your picks perpendicular to the lower-scoring axis. Mind you, you won’t improve your expected value, you’ll only have more multiple-quarter wins and shutouts and fewer single-quarter wins.
I can’t imagine any reason to oppose pre-labelling, unless the idea of somebody else playing as per the previous paragraph offends your sense of equality.
And to give a counter-example, I won my pool’s halftime bet a few years ago when Jerome Harrison “scored” a “touchdown” at the end of the first half of Steelers vs. Cardinals.