Going into game design was the first thing I did when I got out of high school, and while for 2010-2013 I have not been involved with anything that has shipped besides my focused feedback with Blizzard an some personal projects, just this year I picked up a new gig that I hope I can talk about in about four years or so year when we're ready to promote it.
While this isn't related to your post, I'll say it here anyway. The three reasons why I don't put too much stock in the Gematsu leak, aside from the fact it is objectively partial:
1) No game in history has ever gone as planned
There is NO WAY you can ever have all the characters, features, stages, etc. work out as planned. Especially within the first year, things that sounded awesome in paper just don't work as you had hoped. For example, for all we know they could have planned to have Chrom replace Ike, but decided against it. Even the most experienced, most prestigious developers in the industry have to adapt their flow as they move along. It's just a natural part of the process, and there's nothing wrong with it.
2) Sakurai is stubborn, but still volatile
Related to the above. While this isn't related to my experience and knowledge of game design, it is related to Sakurai.
Sakurai claims the roster had been decided from the beginning, and this is true; but it is nigh impossible to be set in stone. Every single one of his games, from Kirby to Smash Bros, has unused content, proof of things removed, things partially added, and he even talks about things that don't work. This is a part of the game design process. The "Forbidden Seven" is a popular example, but it goes way deeper than that. To think Melee could have had Leif and not Roy, Lucas and not Ness. There is no doubt they were on a design document from the start, too.
As stubborn as Sakurai is, even he has had to change his plans all the time. That said, with Brawl he clearly bit more than he could chew, and he is likely avoid it with Smash 4. That is why he would try to have a stricter plan from the beginning and stick to it from the beginning.
3) A lot of people will have information that are not touching the game
As I have mentioned before, even with a top secret project like Smash Bros, a lot of hands will be touching this information. The fact it gets leaked so close to when it is revealed, and the leaker used a term specifically for the English localization, supports the fact it's not anyone involved with the development team themselves.
The first and likely possibility is the leaker is someone who screens the content at e3. Since, you know, it will be seen by thousands in person and millions around the world, hundreds of thousands of those watching it live. The second possibility is someone involved with licensing, since even if you are using an IP that is 100% internal and owned by Nintendo, there is still lots of paperwork involved, which is heavily reflected by all the text you will see under the title screen of a Smash Bros game. And of course the last possibility is someone at any level of marketing. While I doubt someone from Treehouse is a leaker, with such a massive marketing department, Nintendo like any company will send things to marketing when they need them to work their magic. With an IP as big and important as Smash Bros, they will need to get everything they can on top of it to make it good and make it fast.
____
And it's not even like I'm reaching here because I don't want Chrom or something, because all I care about is whether or not the game is good. Whether we get Chrom, Robin, Lucina or who ever, nothing definitively makes them a good character until I have seen them in motion. I have just been around for so long to painfully watch how badly people can blunder in their speculations of Smash 4.
You don't even need to be involved with game design to be able to think about this, really. Even if the leak had said Robin and not Chrom, I wouldn't let it affect my speculation. Critical thinking would not allow me to place any stock in it unless I have seen a working build of the character late (2014) in development, or until the product has shipped.