If I hold another NCAA picks contest next year (as of now, I tentatively plan to, but who can predict the future?), there will need to be some changes. There are a few problems that arose in this year’s contest that I’d like to address next year, not the least of which is the basic format of the thing is flawed. As the pre-bowl standings suggest, the format would allow someone to just choose favorites the entire season and be in contention to win. So, there will be a format change of some kind next year, because people betting that way would make for a boring contest until the last week.
There are a few possibilities of how to do that. Scott suggested betting on up-or-down favorites rather than the spread, but factoring the odds in. I’ll explain how that (and another option) would work momentarily, but first I’ll have to clarify a long-running mistake of mine.
It might irk some of you that what I’ve been referring to as odds are actually the point spread. If X team is favored over Y team by 16 points, that’s a spread, not odds. But I called it odds last year without giving much thought to the nomenclature, and that carried over into this year. Odds are when a team is favored 2:1 or 3:1 or 16:1. So there.
With that out of the way, here are two options I’ve come up with so far as possibilities for how to run the contest next year:
Option 1: Weighted favorites
This is the format Scott suggested. Here would be an example of how to bet on up-or-down favorites, with odds factored in:
Tennessee is playing Kentucky
Tennessee is favored 16:1If Tennessee wins:
Betting on Tennessee was worth 1 point
Betting on Kentucky was worth 0 pointsIf Kentucky wins:
Betting on Tennessee was worth 0 points
Betting on Kentucky was worth 16 points
That would provide a much greater incentive to bet on an underdog without fear of losing your shirt in the standings, and allow at least a high percentage of players to remain in contention through most of the season.
Option 2: Graded spread
The other option I’ve concocted is to bet on the spread, but with a slight twist. Here’s another example:
Tennessee is playing Kentucky
Tennessee is favored by 24 pointsIf Tennessee wins by 24 or more (covers the spread):
Betting on Tennessee to cover the spread was worth 1 point
Betting on Tennessee to win but not cover the spread was worth .5 points
Betting on Kentucky to win was worth 0 pointsIf Tennessee wins by less than 24 (does not cover the spread):
Betting on Tennessee to win but not cover the spread was worth 1 point
Betting on Tennessee to cover the spread was worth .5 points
Betting on Kentucky to win was worth 0 pointsIf Kentucky upsets Tennessee:
Betting on Tennessee to win was worth 0 points
Betting on Kentucky to win was worth 1 point
I think this option would have a feel much more like this year and last year’s contest, while at the same time discouraging people from taking straight-up odds all the time.
Statistically, I have no idea how either contest would turn out. Please provide some feedback, including your own suggestion for a new system if you have one, as the format needs to change next year.
i like the second idea.
Hmmm, interesting. I don’t know that I have really strong feelings about either of them, but I think I prefer the first one for one simple reason: it would be easier to make the picks. I could still get on and say “favorites for everyone but …” and that’d be it. But in the second system I’d have to say “Tennessee will cover the spread, Georgia will win but not cover, Boise State will upset Notre Dame …” and that might be cumbersome.
It sounds interesting. I wonder if it would make it easier if everyone submitted their choices on a specific spreadsheet format so that you could simply cut and paste spreadsheets into one giant spreadsheet.
I was guessing more or less wildly, so that tells you what my strategy was. I’m not sure how the new systems would affect me.
Well, I think it’s probably clear where my vote goes. But I’m intrigued by the proposition of mixing favorites & spread. My concern it that it still doesn’t give enough incentive to pick underdogs.
I’m also biased against spreads because picking against them isn’t really like betting on the underdog - it’s betting on the intangibles, like how generous is the winning coach feeling, and does the winning quarterback have a sore elbow he’d prefer to rest, and will they give the second-string kicker some playing time in a game that’s not close. I’ve seen so many games where the winning team could have beaten the spread but chose not to.
If you want to test odds - is there any way of going back 16 weeks to get the game odds and plugging them in to actual picks? (Can you tell I’m an engineer with time on my hands?)
I’m going with Scott’s plan.
Don’t you mean “…to lose was worth 0 points”?
I am in favor of incorporating the point spread, if only because it is more realistically tangible than odds.
There’s also a much simpler way to do it than with all this half-point business. Let’s say Tennessee is favored by 24 points over Kentucky. You either pick Tennessee or Kentucky. If you pick Tennessee and they don’t win by at least 25 (like this year), you get a big fat zero. If you pick Kentucky, they can lose by three whole touchdowns and you still get a point. You pick a team and they either cover the spread or they don’t. You’re either with us or against us. Stay the course. Marines don’t cut and/or run.
If the final score reflects the exact spread, no points are awarded.
I am going with Scrivener on this and say that the first option makes things a lot simpler, especially for you, Rusty. Keeping track of all the “i’ll take the favorite here, but here only take win but not the spread, etc.” might get really confusing. But since a lot of my picks were going on guesswork as well, I probably shouldn’t speak up.
Garrett,
If you bet on Kentucky to lose, you were betting on Tennessee to win.
Alyssa,
I’m thinking about writing software to automate the whole thing, so it’s about the same effort either way.
Tony,
Yes, that should be considered yet another option.
Scott,
I could test your option that way, but not either of the spread options. Of course, I couldn’t attest to the accuracy of the results since people would likely bet differently under a different format. The better way to test formats would be to do some dummy contests between now and next year to see how they play out. If I write software, I’ll have to beta test it, so that would be one way to find out.
Rob,
You weren’t far off from what I was thinking.
I like option 1. It’s simple and gives incentive to making correct upsets instead of just going with favorites for the most part.
I like option 1 as well.
Merry Christmas, Happy Hanukkah, blessed Eid, a glorious occultation of Spica by the Moon on Sunday morning about ~5:30, and Happy New Year!
I would like to see the straight-up or confidence style pool.