zingy is not the top scorer anymore for this quarter, as of last Weekend. What a turn of events in just 1 week.

Awesome question, and a scary one.

- I looked at the data the quarterly chart comes from, it all checks out the same as the chart.
- I went a level lower, pulled all the aggregated weekly data, and it all aggregates to the same data as in the quarterly chart.
- I then went another level lower, pulling each players predictions for each movie, and the actual for each movie, calculating their score using the formula by hand. It checked out the same as in the quarterly chart.

What this means is I'm confident in their overall scores being correct. I still haven't solved the mystery of WHY or HOW the holdover / opener scores can be higher, because in my mind it shouldn't be possible. If anyone wants to take a look at each players data, you can view it here, maybe you'll have time to do some calculating (I'm at work).

https://docs.google.com/spreadsheet/ccc ... GZSLWFhMVE

There's something wrong with the individual calcs for holdovers and openers - neither is correct. The overall is correct though.

For example zingy's openers are 81.652 and i.hope's are 80.148. Zingy's holdover's are 87.695 and i.hope's are 88.118.

Can you explain the method for your holdover calc, or add a cell on the spreadsheet with the calculation? Then I'll dig up how the calc is currently done.

Damnit.

If you're using excel you would do something like =AVERAGEIFS(H:H,B:B,11,D:D,2) for user 11 openers. So that averages column H with the criteria that column B is user 11 and column D is type 2.

I think right now you might averaging the openers and holdovers by game separately, and then averaging those averages (although when I do that the numbers come out slightly different than what the site has).

I found the error, basically there was an error in our calculation that was treating each weeks holdovers / openers with the same weight (as you suspected). My guess is that the reason it was a little off for you is that some weeks had 10 movies, others 11, and the weeks with 11 would have been weighted slightly more than those with only 10.

In any event, when this current week has actuals added, the quarterly chart will reload itself and calculate correctly.

Thanks everyone for reporting this and helping troubleshoot! Glad we got it fixed moving forward.

Mentioned in the other related thread, but will do so here as well. For this weekend, given the weakness of the bottom holdovers and the intrigue and potential of The Hunger Games, I decided to predict its opening day as well instead of predicting some movie in $1m range. Hope that's good with everyone.

Cool by me - fun deviation.

That sounds reasonable.

VICTORY IS SWEET

Nice job!

And we're back on top!

So the calculations of individual openers/holdovers were done on a per-film basis instead of a per-game basis. Am I understanding it correctly?

Another question: what is used to figure out the relative score, the mean or the median?

I would prefer the median because it mitigates the effect of outliers.

i.hope,

The calculation was being done on a per GAME basis, which is incorrect. Each individual prediction should be weighted the same. Calculating it based on a per game basis was causing some predictions to be weighted heavier than others.

The relative score is currently calculated off the derby average, aka the mean. For now I haven't noticed any issues with it. If we do, we can certainly re-visit the calculation.

Apparently an easy weekend to predict. 25% players have a score over 90% based on the estimates.

Casa is going to end up deciding much of the order. Where did the 1.2M estimate come from? That seems low...

That's the movie that is killing me this week. I hope the actual is a bit closer to 2M.

Seconded. The chances don't look too good though.

For now I just added the estimates for casa from whatever was updated yesterday. Nikki doesn't think it will be in top 10 apparently

Casa de mi Padre - 1.1M Crap

0.08 separates the top three going into the final week.

Pretty high scoring game - congrats bl1222 on the dominance.

I somehow thought Project X would get hit harder by Hunger Games.