I'm interested in comparing the learning rates of people using Skritter to learn new characters vs. those same people using Anki or flashcards or some other method. I think it might be fun to do an experiment: volunteers take 40 characters they don't know, randomly split half of them into a Skritter list and half into a non-Skritter list, and then track how much time it takes to learn them (just the writing).

I'm thinking you would just test yourself after a month to see how many you remembered, with the idea being to remember most of them (so you would have to do some reviews after initially learning them).

I would only need a handful of volunteers to get a rough idea of the learning rate differences between various approaches. Not aiming for full scientific validity here, but rather a ballpark figure.

Anyone interested? Experimental suggestions?

If the Anki prompt was giving the definition like Skritter does, and one was then writing the character out by hand on paper, then the difference would of course be fairly minimal. But far less convenient.

If one was just looking at the character though, then yes the difference would be huge. I add sentences with new words and characters into Anki (in example sentences) before I study them on Skritter. I've noticed I'm able to write some new characters out for the first time after seeing them in flashcards only and having never written them before. But, that's only for less unusual characters, and maybe only 1 in 8. Perhaps over half I have an inkling how to write, but it's not enough. Even after a few weeks of seeing them in flashcards.

If I was actually focusing on how to write the characters from seeing them in flashcards (rather than focusing on using them in sentences/reading them) then I think the ability to write the characters would go up a bit. But it's never going to reach anything like the level of acquisition Skritter would have.

As for your experiment itself, the 20 characters in Skritter, if I added them over 2 days, I'd be comfortable with them in maybe a couple of weeks (if I get them again at the 2 week reappear point, they're probably fairly learnt for the time being). 20 characters in flashcards, just by looking at them - it would take a very long time indeed.

Well, you wouldn't just "look" at them--you'd use the flashcard or Anki to prompt with the definition, and then you would write by hand or in air.

My hypothesis is that Skritter will actually cause them to be learned faster than writing them out from Anki, in contrast to your guess that the difference would be fairly minimal.

Getting Anki up to a level where I'd want to use it for writing characters would be an absolute pain, and it would never be nearly as convenient or well done as Skritter. But at a basic level, they would both be SRS systems prompting you to write a character.

Apart from the innumerable conveniences Skritter has, in terms of memorisation the advantages I can see are: in a mnemonics system (if you use it - although even this could be set up in Anki for someone who can code), character breakdowns (if you use them), and the stroke order (if you need it). I have the impression Skritter's scheduling may also be moderately better. On the other hand, in this scenario Anki has the advantage of actually writing the character out cold, without any hints at all.

It's impossible for Anki (or another flashcards program) to come close to Skritter for the ease in learning writing characters, but a basic level, I can't see how the memorisation would be that different if one spent the time to set Anki up well, and could deal with the pain of doing so. (In that sense, I suppose, they would be learnt faster on Skritter).