OK, I'm trying a new approach, given that my retention rate has dropped a bit and I seem to have quite a large backlog of old items to review (around 500-700 a day). As a result, I've paused all my lists, so no new items are being added, and I'm going to spend the next few weeks just going through what remains. When I have more thoroughly mastered the backlog of tricky items, I'll start adding some new items again.

I suspect that when my backlog gets too large, there's simply too much old stuff to remember to have new stuff added, and so the time spent learning is not too effective.

Anyone else tried this approach? There's been some discussion on the forums about making the SRS algorithms more intelligent; but I wonder if what I personally need is to make my own approach to using Skritter more intelligent... And this seems like a step in the right direction.

That's essentially what I did after I got back from my Colorado ski trip and 10 days of no practicing. I think I had nearly 400 items to review and because all the review times were scrunched together and I was supposed to have reviewed way before I did, I couldn't remember a lot of the items and my retention rate took a huge hit.

Now that I've worked my way through the backlog, the scheduling is working once again, but it was painful getting through. It sounds like you are a great deal more ambitious than me in terms of adding characters, so I imagine your up against an even harder challenge.

The backlog a lot of users are dealing with has been caused by our huge update and a few of the bugs along the way, but it should get better. Your idea of pausing lists and working through the backlog is a good way to get back to a more manageable practice schedule.

Alas... too much ambition too early on leads to indigestion later. But I'm happier now that I've got no new stuff coming in, and can just concentrate for the next few weeks on what I've already seen.