As we come to redesigning the add word frequency option, we're thinking that it's not very easy to understand. Even if you get that it controls how early in your review session the kitten starts adding new words, you still don't know how early each option corresponds to until you've tried them, which takes a while.

One option we were considering includes a slider paired to a review bar, such that you could choose how many prompts remaining you'd have before you'd start seeing new things. You'd also be able to slide it all the way to 0, or choose "manual" such that it wouldn't add anything unless you told it to.

But even clearer would to just have two settings: "automatic" and "manual". "Manual" would be the equivalent of pausing all your lists, and "automatic" would be our current default add word frequency. This option would be on the practice page next to the active lists display, so it'd be easy to change.

You also will now have the "add word" button sitting next to the "added" bar, so it's easy to manually add a word whenever you like. It seems to us that that's a more straightforward and flexible way to control new item flow than a choice of different automatic addition speeds.

We checked the data, and 191 of you have changed your add word frequency. We can't easily tell how important it is to those of you who did, though, or how much you would want that setting if you add a "manual" ("pause all lists") toggle and a manual add button.

So that's our best idea. Please let us know if that wouldn't work for you, if you love the add word frequency control, if you had no idea what it was anyway, or if you have a special use case to be considered.

Thanks!

Ahhh, I like that! Would love to be the decision maker on when words are added, and even which words! Sounds amazing!

I'm not sure exactly what I'd like. It seems that if there are characters that I "just don't get" because they have no pattern that I can relate to (examples are 橘子，酸） then the scheduler repeats it until I can execute it competently and then seems to think that the character is "learned", whereas one week on and I again have trouble.

I've ended up putting particularly troublesome characters into a cram list.

Another thing I might like is that the strictness should vary depending upon how much the character has been practiced. I'm not sure skritter's definition of "strictness" really captures this. For example,
if a 竖 fails to cross a 横 when it should, should be classed as wrong for a character that is well-learned.

Otherwise, the current setting seems fine and is delivering what looks like good progress what is supposed to be the ideal retention rate.

This is one feature that I thought I would "get back to" after I got used to the system. At the start I didn't know what to make of it, mostly because I didn't know how I would do at the whole Skritter thing. Turned out that while Skritter has been a huge help (thanks)ibn learning I never did get back to that feature. My only control over adding words now is to pause my lists when I feel like I am not ready for more. Sometimes that's because I feel like I am overwhelmed with new characters and not really learning any of them; sometimes I feel like I want to practice but coast a little. I suppose that's not ideal but it is reality. It sounds like you're formalizing just what I (and others) do already. Fine!

And, like 阿福, I have some troublesome words （I just cannot get 骄傲). Thanks for the idea of making a cram list of them.

I say nuke the current vague categories and implement the system you've mentioned. As long as you can pause all the lists, and alter the add word frequency when it's on, I'm happy. I alter the settings to catch up with words I have to review, and to speed up the frequency at times.

I agree that the most common thing I do is pause all the lists. Other than that I have just changed the frequency to try to see the difference at times. I think at most you might need a "default" setting and one "add faster" setting, just because if someone found that they could handle words more quickly, having to manually hit add all the time seems a bit overkill. Then again, I don't anticipate myself having that problem, and maybe it would be worth it to have one less setting to worry about.

I am very excited about the add word button but a little sad about having my add word frequency just set to a default setting. I have it as high as it will go right now and I love it. I still feel like it doesn't add frequently enough sometimes, so I guess I would be using the manual add word button quite a bit.

@sarac - when I run into a troublesome character, I find as many two- or three-character words using that character as I can, and I put them all into my queue. I do the same thing with any unfamiliar components of the character, using MDBG's component breakdown.

This means that I am repeating the troublesome character at a higher frequency as part of those words, and with the additional context provided by the containing words. This never fails, although it does slow down my overall rate of adding new characters.

James

With reference to @jww1066's comment above, I think this should be a feature that the app can perform automatically!

I like having the system automatically adding words. I'll either already know them, or know the sounds but not the characters, or it will be entirely new and I'll check for examples on other web sites. As to the frequency, I'd say that should be relative to an individual's retention rate, right? If someone only has 50% retention on a non-trivial set, perhaps all review is in order.

Having both the "default" frequency and the manual adding option would take away my need for other frequency options. There's always the "stop" button in cases where the default setting is adding more than I want.

The only time I change the word frequency is in moments of panic; but pausing the lists works just as well, so I don't think that I'd miss it.

Good tip about adding bunches of words formed of troublesome characters, James. I'll give this a go.

Shouldn’t Skritter be smart enough to adjust the add word frequency according to the deviation of the actual retention rate from the target retention rate, and add more words when the actual retention rate gets high and less when the rate becomes low? Let the users indirectly control the add word frequency by adjusting the target retention rate, and knowing the consequences in doing so, i.e. more reviews to accompany more new words.

Having only one "automatic" setting allows us to try all sorts of clever things to better adapt the adding rate to the user. Matching it to the retention rate might work well, or trying to get an estimate of how much you usually practice, or adapting to how often you hit the "manual add" button. Part of why we want to get away from user-chosen values is that then people have an expectation of how fast that we would hesitate to change.

We may or may not add a "fast" option later, but I think we'll start with just "automatic" and "manual". If it doesn't suit, we can adjust further.

I think the two setting idea that you prefer is fine. Either add new words at a decent clip, or stop for a while. Works for me.

"Even if you get that it controls how early in your review session the kitten starts adding new words,..."

What kitten?

Oh, we often call the "spaced repetition scheduling system" the "kitten", in the same way "squigs" are "strokes you're writing that haven't been recognized yet", the "phantom" is the "character outline you get when you press the show button", the Chef is the program that builds the characters out of component strokes, the Infant is the sound system, the Genius is the word loading system, the Baller is the word editing system, the Drudge was the squig categorization program, etc.

Don't even start on naming conventions inside the code itself. There are Burritos and Politicians and Rocks and Mauls and even a Simon.

One can only do so much to limit the amount of "character stroke recognizers" and "vocab list display" and "item scheduling factor" names, though.