You're missing out the iota under the last letter of the second word. This has value 10, giving 3627.
Because it's a a diacritic, not a letter. If you want to count diacritics that were once derived from letters but weren't letters anymore, that would include the daseia over the "o" three times, for 70 apiece, making the honest total 3837.
I'm not sure why you don't understand the procedure.
Because the phrases you used to identify what values to put in to the calculations are vague.
Jenkins clearly displays the calculations
I was supposed to track down something by a "Jenkins"? You're the one who's seen that before; why not put it here yourself?
wouldn't you say that this is a phenomenon worthy of further study?
What kind of study would you do? What would be the parameters for an accurate prediction and an inaccurate prediction? More calculations to try to find more recognizable numbers would, at this point, become not just noticing something unexpected like it was at first, but a fishing expedition. Once you've derived several more numbers from the one original number you started with, the number of options to create new calculations from them multiplies rapidly. Just pick any combination of some or all of the numbers you've thrown out here so far, pick any mathematical operator or combination/sequence of operators, and start sticking them together in whatever length and order you want like
LEGO bricks... and then ignore the constructions that don't yield something you would want to point out.
If you concatenation the verse values, to give 27013627, then square this number, you this time reveal the first few digits of alpha, the electromagnetic fine structure constant.
No, you reveal a 15-digit number that starts with 4 digits you like followed by 11 you don't because they are wrong for this claim. (You might try to squeeze a fifth coincidental digit out of it, but you'd be rounding that digit down when it actually rounds up.) At least when you were talking about π and
e, without seeing the actual results you were talking about, I could give you the benefit of doubt and presume that they might be accurate up to the point where they were rounded off, but this one doesn't stop at the rounding point: it continues with more stuff that we're now supposed to pretend isn't there. And it starts after the decimal point and two zeroes, which brings back up the decimal system problems I mentioned before, only worse this time. And it's using an input number you only got by treating one diacritic as a letter but not others. And it's the result of one combination of the collection of numbers you have now with a couple of mathematical operators, out of I-don't-even-want-to-guess how many other comparable combinations could have been done but have been ignored.
And why are you trickling this stuff out one at a time? You think that if we weren't convinced before, one more will do the trick next time? This is like a stage show where the mathemagician takes up as much time as possible, acting like he's done one minute and then going "But wait, there's even more! Isn't it amazing how I just keep going & going!" It might be decent theatrics, but it's inefficient communication. And even for just making emotional impressions, it might help if the claims weren't progressively getting weaker as we go along.