How does this work? >
The code works on the principle of polynomial regression:
In MET 2020, board marks were implemented with a 60-40 weightage—60% being MET and 40% board marks. Data compiled from various sources (YouTube, Reddit, Quora, and direct student input) is used to compute band percentages with the following formula:
\( \left( \frac{M}{800} \right) \times 0.6 + (B \times 0.4) \)
Where M stands for your MET mark out of 800, and B is your board percentage. After calculating band marks, these scores are plotted and compared against their ranks.
Using a mathematical concept called polynomial regression, we can estimate the ranks for different band scores. Note that:
However, this remains the best method to predict this year's rank.
Why use the 1 band system instead of 0.5?
ChronocRafter (credit to him) discovered that the data best matched when using the 1 band score rather than 0.5 or a 60-40 average. With other methods, you might see better band scores correlating with worse ranks. For a detailed look at the data, please review the table below:
| MET marks (/800) | MET % | Boards % | 0.5 Band score | 1 Band score | 60-40 avg score | Rank |
|---|---|---|---|---|---|---|
| 613 | 76.625 | 95.8 | 85.975 | 85.975 | 84.295 | 108 |
| 579 | 72.375 | 96.6 | 83.425 | 83.425 | 82.065 | 228 |
| 512 | 64 | 95.5 | 78.4 | 78.4 | 76.6 | 604 |
| 505 | 63.125 | 96.5 | 77.875 | 77.875 | 76.475 | 672 |
| 480 | 60 | 96.66 | 76 | 76 | 74.664 | 950 |
| 406 | 50.75 | 96 | 70.45 | 70.45 | 68.85 | 2234 |
| 490 | 61.25 | 85 | 72.75 | 68.75 | 70.75 | 3016 |
| 382 | 47.75 | 93 | 66.65 | 64.65 | 65.85 | 4493 |
| 375 | 46.875 | 91.5 | 66.125 | 64.125 | 64.725 | 4700 |
| 370 | 46.25 | 93 | 65.75 | 63.75 | 64.95 | 6700 |
| 375 | 46.875 | 81.5 | 62.125 | 56.125 | 60.725 | 11012 |
| 233 | 29.125 | 87.76 | 49.475 | 53.475 | 52.579 | 15762 |
| 427 | 53.375 | 70 | 62.025 | 52.025 | 60.025 | 17044 |
| 356 | 44.5 | 75.5 | 58.7 | 50.7 | 56.9 | 18132 |
| 568 | 71 | 0 | 42.6 | 42.6 | 42.6 | 21000 |
| 240 | 30 | 69 | 46 | 34 | 45.6 | 31500 |