@let's fight: your video does not load for me when I follow that OneDrive link. YouTube works, though!
---
Here's how actual results (boxed borders) compare with predictions from
https://bit.ly/cl14-predictions (usernames are people who predicted that clan would finish at that rank, numbers are counts of people):
Predictors, Rankedbased on: SSE, then SAE, then # of Correct Predictions, then listed in order of when they made their predictions
Rank | Predictor | SSE | SAE | Correct Predictions | r | r^2
---------------------------------------------------------------------------
1 | Ralph | 14 | 8 | Python, MASTER | 0.750 | 0.563
2 | Sandokan | 20 | 8 | Python, ONE!, GG | 0.643 | 0.413
3 | Ursus | 20 | 10 | Python | 0.643 | 0.413
4 | OTTERYCAT | 26 | 10 | ONE!, GG | 0.536 | 0.287
5 | Checkmqte | 26 | 12 | | 0.536 | 0.287
6 | Phobos | 28 | 12 | | 0.500 | 0.250
7 | Aura Guardn | 30 | 10 | Python, ONE!, GG | 0.464 | 0.216
8 | l4v.r0v | 30 | 12 | ONE! | 0.464 | 0.216
8 | Beep Beep | 30 | 12 | ONE! | 0.464 | 0.216
8 | Onoma94 | 30 | 12 | ONE! | 0.464 | 0.216
8 | Hades | 30 | 12 | ONE! | 0.464 | 0.216
12 | Phaeril | 30 | 12 | | 0.464 | 0.216
13 | Samwise | 32 | 12 | GG | 0.429 | 0.184
- | CONSENSUS | 32 | 12 | GG | 0.429 | 0.184
14 | Farah | 34 | 13 | Python, ONE! | 0.393 | 0.154
15 | Meldarion | 34 | 14 | | 0.393 | 0.154
15 | Octane | 34 | 14 | | 0.393 | 0.154
17 | Tukuro | 36 | 14 | | 0.357 | 0.128
18 | Ocean0.1 | 38 | 12 | ONE!, GG | 0.321 | 0.103
19 | let's fight | 38 | 14 | GG | 0.321 | 0.103
19 | We ride! | 38 | 14 | GG | 0.321 | 0.103
21 | Kratt | 40 | 12 | HAWKS, GG | 0.286 | 0.082
22 | Ze | 42 | 14 | MASTER | 0.250 | 0.063
23 | Xenophon | 42 | 16 | | 0.250 | 0.063
24 | Tjoex | 44 | 14 | GG | 0.214 | 0.046
25 | lineal2012 | 46 | 16 | | 0.179 | 0.032
26 | rakleader | 52 | 16 | | 0.071 | 0.005
27 | goralgn | 62 | 14 | Python, HAWKS, SNinja | -0.107 | 0.011
28 | Nature | 66 | 18 | ONE! | -0.179 | 0.032
29 | finding mwc | 74 | 18 | ONE! | -0.321 | 0.103
30 | Norman | 80 | 20 | ONE! | -0.429 | 0.184
SSE = sum of squared errors (e.g., 16 points for guessing FCC would finish 5th); higher is worse
SAE = sum of absolute errors (e.g., 2 points for guessing MASTERS would finish 2nd); higher is worse
r = Pearson's (and in this case, also Spearman's) correlation coefficient; computed as 1-(SSE/56), thanks to Math Wolf for the formula. 1 = perfect, 0 = uninformed, -1 reversed. Above 0.8 is considered good.
r^2 = coefficient of determination; what % of the variation in results aligns with the predictions, can think of as what % of the information about final rankings someone actually had in their prediction. E.g., "Ralph was 9/16ths right, while if you did the opposite of Norman's predictions you would've been 18% right. rakleader's predictions, on the other hand, had basically zero information."
Edited 9/5/2021 13:49:56
