Play with the numbers to see how the success rate of any individual localization team plays together with string freeze cycles, to see how likely we'd ship how many locales.

If you're curious about the algorithms used, check out the annotated source of math.js. Feel free to clone/patch/issue on github.