You start with the dates of all the FIRST cases of a worldwide Direct in that month. 2 Directs in a month is always unrealistic to believe as it doesn't really happen often (not saying it can't!). The spreadsheet automatically analyzes them to find the day of month and the weekday.
Using that, I then discover for each day of the month what are the dates for that day throughout the month and then also out of the pool of data, which ones have the highest percentage. That shows their frequency.
We also find the statistics of the trend. The earliest point a Direct happened as well as the Latest and the averages. It helps give clarity.
The data is then sorted in ranking with the highest percentage being the top pick while the bottom ones being the least likely. You can start to see the patterns when you sort them correctly.
But this is only half the battle. We need to filter it. Remember we looked at the statistics? It's time to apply them to get our realistic guidelines. Firstly we remove anything from a day that has never been used (only Wednesday, Tuesday and Thursday remain)
Next is the important one. As you can see from the pool of data, for the 1st Direct in each month, they have never gone past the 13th. So using that maximum value, you can filter the dates further. Notice how the rumored 1st of November date is there? Remember this for now.
This next step is optional but only for those who want a minimum value. We do the same as maximum but for the minimum too (which is the 2nd)
As you can see, the 1st vanishes! But only by one day. Isn't statistics and data fun?
What does this mean for the 1st? Well it is still plausible. Had the minimum value not been applied, it would have been in the pool of realistic dates for this month.
It was only out by one day! So this rumor has actual solid ground. The best part about using the automated Spreadsheet is that it isn't random guesswork. It is data crunching so it is unbiased and fair. There is another part why and I have put it in bold it later in the explanation.
As for the "2 Directs in a month" rumor, I looked into that too and spotted some information for you all. It has happened before in November.
November 7th 2012 (Wii U Direct) and November 27th 2012 (New Super Mario Bros 2)
November 13th 2013 (General Direct) and November 27th 2012 (Nintendo 3DS Louvre)
There's a pattern here too. It may not look like much but if you expand the search to the other months it is there too. If there is a Game-Specific Direct in a month where a Nintendo Direct is a lock, the chances of two Directs was increased. Not guaranteed but increased (see Xenoblade Chronicles 2). It is that uncertainty that makes me only look at the 1st Direct each time.
But obviously why should this spreadsheet be taken for solidity? Well it has been far more accurate than any "leak" this year or the previous years (I went backwards in the data pool to find out if the spreadsheet had accuracy).
In fact for this year, it only got one date wrong this year and it was out by 1 day. That was the August Smash Direct. The spreadsheet predicted the 7th.
Obviously don't take this for gospel truth. It is an automated spreadsheet so could be wrong! No-one is perfect, But in terms of plausibility of two Directs in a month, this one actually has solid ground.
But it all depends on that November 1st. If it the first Direct doesn't happen by the 13th, you can consider this "rumor" debunked as it makes it very unrealistic to have two in a single month.