Niners Nation Prediction Games: A Call for Help

Hello, you Niners Nation faithful.

Those of you who have been with us for a while already know about our proud, annual Predictions Games contest. For those of you who don't, it's a game that we hold throughout the season. All members of the community can play, and - if we can successfully arrange it - prizes are sometimes even available to the winners. A more full explanation of the game will be visible after the jump, but first: the call for help.

Over the last three years, Fooch and I have been the sole administrators for the games (Fooch, who selects the contests, and me, who scores and reports the results, with the exception of one week, when groug quite nobly held my place). This season, that is not going to be possible. I have more on my plate than ever before, and I just can't see myself reliably performing this function on a week-to-week basis. In short: I need a replacement.

If anybody is willing and able to take my place as the official scorer for the prediction games, it would be a huge boon not only to myself but also to the website at large. I would hate to see something that has been so much fun for each of the last three years die on the vine just because my life got a little busier than it has been to this point (in the immortal words of Blink 182: Well, I guess this is growing up).

After the jump, I'll fully explain the games. Please use the comments to nominate yourself for my position or to ask any questions you have about the process in general.

We call these the Prediction Games because it's an accurate title. Each week during the course of the season, Fooch will select six NFL games. Your job is to predict the winner and the score of each game. It's really that simple. Predicting the correct winner makes you eligible for points, and predicting the score more accurately than others will get you more points. Points accumulate over the entire season, and the games do not end until the last whistle has blown on week 17. Then, we have traditionally held the games again for the playoffs, with some fun variations - but we can worry about all that later.

Let me repeat the important part:

Each week during the course of the season, Fooch will select six NFL games. Your job is to predict the winner and the score of each game. It's really that simple.

If you don't want to, you don't have to worry about anything else. You can ignore the rest of this e-mail for all that it matters. However, if you want to understand the games in more detail, or if you would like to take my place as the official scorer, I encourage you to read on.

To clarify: To play the game, all you have to do is to predict the winner and the score of the game itself. Everything that follows this point is designed to help explain how we score your predictions.

The primary conceit of this game is that any score is made up of nothing but three simple components: 1) Winner/Loser, 2) Point Differential, and 3) Total Points.

To illustrate, let's say that the 49ers are playing the Broncos. Given only three components, you can determine the exact score, with no possible variations. 1) The 49ers win. 2) Point Differential = 45. 3) Total Points = 65. There is only a single possible score that these three things can add up to: 49ers 55, Broncos 10.

Knowing this, we can use these three components to accurately determine how close any given score prediction for any given actual game score is in relation to any other given score prediction for that same actual game.

It works like this: Your prediction has these same three components as the actual game score. All we have to calculate is the difference between each.

1) If you predict the correct winner for an individual game, we will score your prediction against the actual game. If you predict the incorrect winner for an individual game, your prediction will not be scored.

2) Ideally, you will predict the exact point differential of the actual game. Say, for example, this is 45. If you predict a score with a PD of 45, the difference between your predicted PD and the actual PD is 0 (45 - 45 = 0). The worse your prediction, the greater that difference becomes. Therefore, the closer the difference is to zero, the better.

3) Ideally, you will predict the exact point total of the actual game. Say, for example, this is 65. If you predict a score with a PT of 65, the difference between your predicted PT and the actual PT is 0 (65 - 65 = 0). The worse your prediction, the greater that difference becomes. Therefore, the closer the difference is to zero, the better.

So, assuming that you predict the correct winner of the individual game, we simply add the difference between your predicted and the actual PD to the difference between your predicted and the actual PT. The closer this total is to zero, the better your prediction is. A zero exactly will always be a perfect prediction.

That total becomes your subscore for each individual game. These subscores are then ranked, and the five best (including ALL ties) are awarded points. Perfect predictions are awarded 10 points. If the highest subscore for a particular game is not a perfect prediction, it is awarded 5 points. Points then follow in descending order, so that all of the second best predictions get 4 points, the third best all get 3 points, and so on.

These points are then accumulated by each individual player for each individual game, as selected by Fooch, over the course of the entire season. The player with the highest total at the end of the year wins. Easy.

Now, to the point of doing the scoring. I have been doing the scoring, and it usually takes me in the one to two hour range to put an entire results thread together. Scoring is easy; transcribing (or even copy/pasting) is time consuming. All scoring and ranking can be done simply using an Excel (or comparable) spreadsheet, which is awesome. If you take my place, Fooch and I have the spreadsheet that I have been using (he has it in Excel, though I, sadly, no longer do) and can give it to you. Alternately, you can create your own that best suits your own process. The equation that you have to program in to get the subscores is easy:

pPD = Predicted Point Differential
aPD = Actual Point Differential
pPT = Predicted Point Total
aPT = Actual Point Total
x = Subscore

|pPD - aPD| + |pPT - aPT| = x

Then, once that is settled, just use a table order function to order the rankings by x.

That's all.

I think this is the end of the FanPost. Nominations and questions in the comments section, please.