I feel like you intentionally chose an example week where you did well and I did terribly ^_-
I second PDF's concerns, and would add one of my own - it was actually really hard to not make a hefty winning each week, regardless of accuracy. Hopefully we can come up with calculations that balance incentivising people to take part with actually needing to be accurate to get a good pay-out.
By being slightly above the average prediction, you would indeed recieve a positive payout, so while you did make money there were people who were losing money.
We could look at the option to remove negative outliers that wont count towards the average for example don't count the 10% furthest off
EDIT : i'm going to take this week again
the interesting thing here is that Nights, DMC and the xbox 360 had an average of 0% accuracy so even with 10% accuracy on those you would get a payout, with a high bet about and for example 60% you could get a huge payout.
Which is what you probably remember.
The top predictor from this round put 1000vg$ on the xbox 360 and had an accuracy of 78% (vs. average 0% ) and 140 vg$ spread over the other options. His profit of 992 vg$ could just aswell be read, purely based on his xbox360 accuracy and bet amount
Last edited by kirby007 - on 11 February 2018