This page discusses some of the minigame voting results. Basically, I wanted to look at some of the statistics of the scoring, to see what conclusions can be made. Also, I had noticed that many of the scores were very close to each other, and was curious about that. And maybe I have too much free time (NOT!). But let's get started.

First, a plot of all the minigame scores:

sorry, it's a graph

There are a couple things to observe here. First is that most of the scores are very close together, with Splatform way above the pack (and the low scores pretty far below). Second is that entries 2-42 or so seem to lie on one line, while entries 42-57 or so lie on a second line.

Now let's look a little more closely at entries 2-41. Here is a plot of the scores and the _difference_ between scores (i.e. score(n+1) - score(n)):

I use lynx too, but ya gotta get some gfx.

As you can see, most of the scores are quite close to each other -- the mean difference is just .04 points, and over half the scores differ by less than that.

Now, most entries got somewhere around 30-40 votes. I started to wonder what would the outcome have been if one person changed their score by just one point. If they added a point, that would add somewhere from .025 to .03 to the score -- and since the _mean_ difference is just .04, that could easily have changed the placement.

So, here is a graph of the normalized score differences. What it tells you is, "if someone had given me N more points, I would have advanced in placement". (To get two entries ahead, you can add two numbers, etc.):

Today just isn't your day!

For example, if two people had upped their score of entry 25 by just one point, it would have advanced in rank. As you can see, just one or two points difference would have changed the positions of a lot of entries! In the worst case (my program RAD -- naturally :) just four points, out of 40 voters, would have advanced it (and lowering just one score by a point would have moved it to 13th place!)

So, what can we conclude from this? First, Splatform was a huge winner -- way off the curve, no doubts about it, etc. Second, the voting system is very subjective and very sensitive to changes; said another way, don't worry too much if you got placed lower than you wanted to, because the scores are all extremely close to each other (the differences are down in the noise). And don't worry about trying to "fix" it :).

As with any sensitive system, it would be pretty easy to 'rig' a vote, by getting all your friends to vote for (or against) certain entries. But surely there is honor among minigame participants :).

And finally, if you scored in the top 40 or so, you did very well! Heck, if you took the time to code something up you're pretty ruling in my book! The voting is for fun and prestige (and motivation), so good job to all.

And remember: wait'll next year! :)

-S