Jeopardy! 538 brain teaser

get-averages.js

const row200 = [200, 200, 200, 200, 200, 200];
const row400 = [400, 400, 400, 400, 400, 400];
const row600 = [600, 600, 600, 600, 600, 600];
const row800 = [800, 800, 800, 800, 800, 800];
const row1000 = [1000, 1000, 1000, 1000, 1000, 1000];

const lowest = [row200, row400, row600, row800, row1000].flat();
const highest = [row1000, row800, row600, row400, row200].flat();

const sum = things => things.reduce((summa, item) => summa + item, 0);
const average = things => sum(things) / things.length;

const getAllScores = configuration =>
  configuration.map((score, index) => {
    const preSum = sum(configuration.slice(0, index));

    const doubleScore = preSum < 1000 ? 1000 : preSum;

    const remainder = sum(configuration.slice(index + 1, configuration.length));

    return preSum + doubleScore + remainder;
  });

const low = getAllScores(lowest);
const high = getAllScores(highest);

console.log('lowest average score configuration: ', average(low));
console.log('highest average score configuration: ', average(high));
console.log('average of those two (random average): ', average([average(low), average(high)]));

Although, my assumption that the relationship between the highest edge case and the lowest edge case is linear is just that... an assumption.

If, however, a random distribution of configurations produces a nonlinear relationship, the bonus answer will be inaccurate, but for guesstimates, halfway between edge cases is good enough. Especially when the best and worse case scenarios are so similar.

https://fivethirtyeight.com/features/can-you-pass-the-cranberry-sauce/
awesome! I had it right on the correct submission, any my average fo the averages was less than 20$ off the fully correct solution, the one that would have required 4.1x10^19 runs (or a monte carlo) :D

quite happy with that level of detail. (26166 estimated vs 26150)