Week 3 saw my team take 3rd place in the Yahoo Pick ‘Em league. It makes me feel better that my dad took first place based off two upset picks, Bengals over Packers and KC over Philly. When I asked my dad if he used my strategy, he proudly said he went by gut feel. It also helped that he’s a die hard Cowboys fan and picked them for 14 points.
Here are the results of the first few places, my team highlighted in yellow:
Week 3 had two contrarian picks candidates, both favored teams and ~60% win probability (according to Moneyline numbers) and only 18% of the population picking them. They were Carolina over the Giants, and Miami over Atlanta. Both Carolina and Miami turned out to be winners (guess those folks in Vegas know what they’re doing). There were 6 true upsets (i.e. Vegas upsets) with the biggest one being Indianapolis over SF (10.5 point underdogs). However, most people in the league chose SF between 12-14 points, so it was a wash. That’s the beauty of this method; if there is a big upset, it usually doesn’t hurt.
This week, instead of going “all in” on Carolina or Miami, I decided to go with straight win probability sorted by buckets because I felt like there was enough contrarian-ness going on with both teams so I hedged a bit. If I would’ve gone “all in” on Carolina, like I was contemplating, I would’ve netted 96 points, still 3rd place:
The only way I could’ve taken first place was to go “all-in” on both Carolina AND Miami, scoring 100 points:
My question for myself is, was not going “all-in” the right play? Would going 16 points on Carolina helped at all? With my current picks, if Atlanta would’ve won instead of the game ending with a Miami TD with 38 second left, I would’ve ended up in 9th place… out of the money. If I had picked Carolina all-in with 16 AND Atlanta would’ve won, I’d be tied for 4th with 92 points. Hard to tell, doing one all-in pick helps hedge my Atlanta vs Miami play, but there wasn’t any upside to get me to first place. Food for thought… one day I’ll crunch all the numbers and learn some statistics and have a better answer.