Win Prizes for Making Forecasts on SciCast

SciCast is running a special!

For four weeks, you can win prizes on some days of the week:

On Wednesdays, win a badge for your profile.
On Fridays, win a $25 Amazon Gift Card.
On Tuesdays, win both a badge and a $25 Amazon Gift Card.

swag_he

On each prize day up to 60 valid forecasts and comments made that day will be randomly selected to win (total limit of $575 per person).

Be sure to use SciCast from May 26 to June 20!

27 thoughts on “Win Prizes for Making Forecasts on SciCast”

sflicht May 21, 2014 at 1:05 am

This seems to incentivize certain forecasting styles over others. To wit, making many small edits is more likely to win a prize than making fewer, larger edits, correct? Also, a bad actor might make arbitrarily many +1%, -1% edits in succession to increase her probability of a prize without losing any points…

Reply ↓

Jessie Jury May 21, 2014 at 6:12 pm

@sflicht: You are correct this is not an ideal incentive, but we will be studying it closely to model and make a case for future ones. Yes, it favors several small edits over one large edit. By itself that’s fine. It is sometimes better strategy to move it a little and see if someone responds, rather than move it all the way at once. It is of course easily gamed. We are betting that (a) most of our forecasters will be more cooperative, and (b) egregious misbehavior is relatively easily detected.

0

Reply ↓
Ted Sanders May 21, 2014 at 2:58 pm

Interesting, if you a draft a comment, the comment draft shows up under ALL replies, not just the one you drafted it to.

0

Reply ↓

Pingback: Overcoming Bias : SciCast Pays Out Big!

sflicht May 21, 2014 at 3:24 am

My suggestion: make the play money worth something. Rewards should go to the person with the highest score at the end of the reward period. Yes, HAL has an advantage but he earned it. (With a lower capital allocation than is currently available, too!) Or reward the top N people using a specified scoring function. But rewards for activity do not suitably couple the reward to the forecasting performance.

Reply ↓

Ted Sanders May 21, 2014 at 2:57 pm

I like the idea, but even that incentive scheme (along with the College Bowl) distorts things. There’s a big incentive for me to bet all my money on a random 5:1 payout. It takes no effort on my part and gives me a fair shot at a prize.

0

Reply ↓
Jessie Jury May 21, 2014 at 6:14 pm

Yes, we want to move towards something like this, and have a pilot test now - the College Bowl.

0

Reply ↓
1. dvasya May 21, 2014 at 4:29 pm
  
  Has any team reached 50 people yet?
  
  0
  
  Reply ↓
  1. Jessie Jury May 21, 2014 at 4:52 pm
    
    @dvasya The Leaderboard will be posted tonight with the Team Leaderboard coming later this week with all of those details.
    
    0
    
    Reply ↓
  2. Jessie Jury May 21, 2014 at 4:53 pm
    
    http://blog.scicast.org/2014/05/21/college-bowl-leaderboard/
    
    0
    
    Reply ↓

Ted Sanders May 21, 2014 at 4:06 am

Are these days according to Eastern Time or the time zone of the user?

Reply ↓

Ted Sanders May 21, 2014 at 4:06 am

(And why does this comment say the time is 4:06 am? It’s 9:06 pm here and 12:06 am EST. Hmm…)

0

Reply ↓
1. Jessie Jury May 21, 2014 at 2:21 pm
  
  Thanks for pointing this out, Ted. This should be set to EST and it should now be fixed!
  
  0
  
  Reply ↓

Jack May 27, 2014 at 2:34 pm

When do we find out if we won?

Reply ↓

Jessie Jury May 30, 2014 at 1:47 pm

Winners are receiving email messages indicating they have won. We’re also publishing user names of winners here on the blog.

0

Reply ↓