So it turns out the Hildy has the lowest standard deviation this year!
I guess that means the people who played it all felt more or less the same way about it. That’s cool, I suppose. I’m choosing to celebrate it at any rate!
But it got me to thinking. If we had an award for the lowest standard deviation, what would the trophy be?
A golden banana of discord is kind of fitting. A banana has two opposite ends that will never meet.
So what would the opposite of that be? Something with no ends, maybe? And what color would it be?
The Silver Cheerio of Harmony?
The Rainbow Oreo of Unanimity?
The Glazed Donut of Consensus?
Yes, this is what I’m thinking about this evening.
I thought of this too, and was actually going to suggest it on the survey, but worried that if the game with the lowest standard deviation in a particular year also placed very low, it might risk feeling like adding insult to injury.
You can always add it to the survey (and even include your concerns). The committee take suggestions into consideration, discuss them, and if the idea holds… implement them!
(and then we wait for feedback to see if it was truly a good idea and adjust if necessary)
I can understand that, but Golden Banana by math can’t be first place nor last place and usually lands in the middle, and the achievement is kind of a cool thing.
I was going to snarkily reply that the award should be “IFComp 1st Place” but then realized low standard deviation could be any score.
I can see how it might be a bit of a barb since it’s pointing out consensus by peers and it’s a dice roll what the consensus is.
That would cause the Golden Banana to explode into flames and burn down an important building most likely - since that type of voting scatter represents the opposite of discord.
Theoretically possible, but would require an unplanned Grand Concurrence among so many diverse people that it’s probably logistically impossible.
It’s very rare for the game with the lowest standard deviation to place highly. Aotearoa won 1st place in 2010, Hildy won 7th in 2024 and Tea Ceremony won 10th in 2014, but no other highest-consensus game has ever cracked the top ten.
Indeed, where the consensus is strong, it generally tends to indicate that the game is clearly bad, like Ninja II, which came in last place in 2005. (It’s the highest consensus game of all time.)
For this reason, I suspect that the games of lowest standard deviation will usually not be as long remembered, unlike the Golden Banana games, which are frequently “interesting” in how controversial they are, and thereby have more historical influence.
EDIT: BTW, generating results like this is easier than ever, now that IFComp has a JSON API.
EDIT 2: Added “out of #” to the placement, so you can see which games came in last place or close to it. There are eight years where the game with the lowest standard deviation came in dead last, plus eight more years where it was “nearly last.”
I started this thread as a joke and I’ve had a good laugh at everyone’s suggestions, but I agree that such an award would likely be hurtful to an author who felt his or her game underperformed. And we should avoid that.
I was very happy with how my first entry (Diddlebucker!) did. It ranked in the middle, but for a first attempt, I was pleased with that, especially since some people really enjoyed it. But had my game also won a low SD award, it would have felt like the award was pointing out that the people who enjoyed it the most were outliers who didn’t impact the score much.
So I don’t actually think we should actually go through with such an award, but I have been enjoying the suggestions that people have had and the mathematics discussion as well.
I also found it interesting that the it’s been rare for the lowest standard deviation game to end up in the top ten. I wouldn’t have thought that to be the case, so I’m pleased that Hildy managed to be one of the few.
Based on the likelihood that this would frequently shine a spotlight that games that everybody agreed were astoundingly average/poor I agree it’s probably not a good idea to actually implement, but as a name for the award…
The Iron Potato of Accord
I feel as though something no-nonsense, solid and dependable would be a good fit for a game that everybody looked at and said “Oh, yeah. I know how I’m rating this.”
I definitely riffed on the Golden Banana with the Golden Grapes of Concord (plus, a pun! CONCORD grapes! Don’t worry folks I’m here all night) but equally definitely did not intend it to be an opposite in the sense discussed here.
Beget of Balance makes me think of a very flat distribution. e.g. given 10 votes on a 10-point scale, the ideal balance result would be one vote of each score… compare 5 1s and 5 10s as the ideal polarized distribution and all scores the same for the ideal consensus distribution… And while a flat distribution and the polarized distribution would have the same mean, whould they have different SDs?
I like golden grapes of concord since concord is both an antenym for discord and a variety of grape, but given the symbolic meaning of gold, perhaps it should be reserved for a game that both has little variation in score and a high composite score. E.g.again, using a 10 point scale, the ideal golden grape distribution would be all 8s, all 9s, or all 10s. Perhaps it could be calculated as arithmetic mean of individual scores divided by standard deviation(my first thought was subtract SD from mean, but that might make mean overpower SD… though division might make SD overpower mean if SD can be less than 1(I never did get a good grasp of SDs and how to calculate them from my Statistics classes in HS and at Uni) or lowest standard deviation among the top tenth or top fifth of entries. And then you could have silver, bronze, iron and rust versions forthe second top fifth, middle fifth, etc.
I also find myself wondering if using different means such as geometric or harmonic means could reveal interesting statistics…
