Overall, a decent performance from the crowd this week both straight up and against the total. The crowd is still behind in the spread portion, and we expect that as we get more users to contribute predictions, this number will improve.
CPR Week 2 Summary
- Straight-up: 12-4
- Against the Spread: 5-11
- Over/Under: 10-6
CPR 2017 Overall
- Straight-up: 20-11, 65%
- Against the Spread: 10-21, 32%
- Over/Under: 19-12, 61%
The main challenge is still the total number of predictions. We’re still hovering around 5 but are hopeful that success breeds success. Nevertheless, the over/under number is leaves room for optimism.
A One-off Crowdsourcing Exercise
Cynthia Frelund (@cfrelund), who is an expert in statistical analysis, posted a poll on Twitter and solicited scores for the Week 2 Monday Night Football game between the Lions and the Giants. I did a quick aggregation of 50 predictions and the result was Lions 23, Giants 19. The crowd went 3-for-3.
There were some extreme outliers, but even with predictions such as 13-10 or 16-10, the prediction coalesced after about 30 or so.
So crowd, we need to get to 30! Tell your friends! 🙂
Crowd-wisest Prediction of the Week
This is a new feature we’ll be running every week, highlighting the game(s) that the crowd was the most prescient in predicting. The primary qualification is that the crowd has to get the winner, the spread, and the total results correct. After that, it comes down to accuracy.
This week’s game was the Cardinals at Colts.
|Spread||Cardinals by 2.25 (Cardinals favored by 7.5)||Cardinals by 3|
|Over/Under||32 (under 44)||29|
We don’t measure success specifically on this kind of accuracy, but it’s always interesting to note.
The larger takeaway is that the deltas between the lines, the predictions, and the results. The Cardinals were favored by 8.5 points, and the over/under line was 44. We believe that, over time, the crowd can identify the lines where the perception of the teams is significantly different from the actual quality, and this is a great example of that.
Crowd-foolishest Prediction of the Week
Of course, for every two wise predictions, we figure that there will be one foolish prediction.
There were two bagel games: Vikings at Pittsburgh and Washington at Rams. I’m discounting Vikings at Steelers because of the late-breaking news that Vikings QB Sam Bradford was not going to play.
|Spread||Rams by 4.5 (favored by 2.5)||Washington by 7|
|Over/Under||38.5 (under 46)||47|
Clearly, the Rams’ big win over the Colts and Washington’s struggles against the Eagles gave people the impression of a bigger discrepancy in quality than there actually was. In the end, the game was quite close and was only decided in the last few minutes so it will be interesting to see if there is an indicator when the crowd believes a game will be close one way or the other.
Top Performers This Week
Big week by user B.C. who topped all three leaderboards!
- B.C. (13)
- C.A. (11)
- P.C. (10)
Against the Spread
- B.C. (10)
- P.C. (8)
- C.A. (7)
- B.C. (9)
- G.A. (8)
- P.C. (8)
Thanks for all the predictions! Keep up the good work!