Hi, I'm interested in how this figure is calculated, and whether it is calculated daily or the figure is real-time.

Mine seems to fluctuate in ways that I wouldn't expect. Some days it leaps very high and others it is quite flat.
And other days it goes down :(

We equate how well you know a character or word or tone with the length of time between reviews. So if a given character has a week between its last review and its next scheduled review, it's known better than another character with only an hour between the last and next review. The better you know something, the less often you need to review it.

We decided 12 hours would be the (somewhat arbitrary) time delineating knowing and not knowing. So basically if you are highly likely to remember something after 12 or more hours of not reviewing it, then you have 'learned' it. What do you guys think, too high or too low a marker? Unfortunately, it's somewhat difficult to change it at this point; we'd probably have to pull the site down temporarily and do some number crunching if we were to change it so that everyone's progress would still be accurate.

The statistic is updated with every review you do. So if you learn a character, you'll see it on the progress page immediately. But it also keeps track of which ones you knew then forgot! So if you forget more than you learn your stats will go down... :( That tends to happen to me when I haven't studied in too long.

I think it works well. I can clearly see progress when I'm sticking to my minimum daily bite size and likewise decline when I have to skip the "Skritter class".

Thanks for the answer, the figure certainly is interesting and gives a good indication of short term progress.

However I certainly don't feel comfortable answering the question "How many characters do you know?" with my skritter figure.

The 6 day review is a killer!

Using skritter has made me realise more and more that revision is a continual and time consuming progress, and you can never really over-study anything, because I WILL forget; its just a matter of how long until it happens.

I don't agree with the statement that you will forget no matter how much you study. The spaced repetition studies make it stick longer, and if studied enough (with the correct repetitions) - forever.

Just don't add too many characters at the same time, then the repetition tide will hit you.

You need to find that correct balance of how many new chars that you will "accept" to learn on every session. This is very individual based upon both the number of sessions per week you are willing to spend, how long each time, and your overall ability to remember, of course.

Try to make a session every day to keep the "waiting number" as close to zero as possible. If you have time, accept new words. After passing the "zero", it is up to you how many new chars to want to add to "the queue", i.e. future reviews.

I think the other side of remembering characters is actually using them. Skritter seems amazingly efficient at drilling the structure of characters into the memory, but if characters are only seen in isolation as individual things to be remembered, it is harder to recollect them. However, if you keep coming across them in a range of contexts, then they take hold much more deeply. I think there's an interplay between form, sound and meaning.

ThinkBhudda, I totally agree, I was doing translation work yesterday, and the characters and words that I accessed in context were solidified more efficiently than if I had been doing rote practice (even with Skritter). I find that Skritter isn't able to link the English definition and the characters as well as analogue practice. We didn't really anticipate that when we first designed Skritter, and it will be included in the big update that's on the way. Our insistence that the prompt system be more comprehensive added a lot of complexity, which is making Nick and Scott's job hard!