掲示板 Forums - Daily study goals mis-accounts for review limits

For a sentence schedule without a limit on daily reviews, Daily study goals shows:

Setting a daily review limit of 10 questions, however, results in:

I think this is weird: it clearly tries to respect the daily limit, but then keeps the 0s while bumping the 8/7/8 up to 10. I would've expected 10/6/10/4/8/7/8 instead, matching the number of questions I would be asked to review each day assuming I stick to the study goals.

I also think the way it handles today's daily review limit is a bit inconsistent: over-limit terms up for review today will be added to tomorrow's daily study goal, but only after you've reached today's limit.

It would be very cool (but probably too much of a pain to implement!) if this projection assumed I was going to do the reviews up to the daily limit, didn't get terms wrong, and accounted for additional near-future reviews of low-mastery terms, so it could serve as a more accurate early warning of days where I'm going to end up with a large number of stacked reviews. The current projection is accurate for tomorrow (after you've done today's reviews!), and then consistently underestimates the review load for the following days.

This is a visual issue in the sense that actual review scheduling behaves correctly, but the daily study goals preview is incorrect in its predictions. I don't think this is connected to switching review limits: the issue persists reliably long after the switch.

One example from today: I'm done with today's goal, and have two ready-to-review terms left over. Daily Study Goals forecasts:

... 0 reviews tomorrow. I will actually have to review 2 terms -- the left-over ones from today. My psychic debugging powers* tell me that only 8 terms will become ready-to-review on 6/27: Daily Study Goals just pushed the two left-over reviews from today onto the first non-zero review day in the future.

I think the observed behavior is consistent with iterating over a SELECT next_review_date, COUNT(*) FROM schedule_terms GROUP BY next_review_date ASC result set, keeping track of the number of "left-over" reviews, and increasing or decreasing the counts to match the daily review limit. On days with 0 terms becoming ready-to-review, there is no corresponding row in the result set, so the left-overs are pushed further into the future than they should be. Is it possible that the code rendering the Daily study goals box does something along these lines?

(* a duplicate of that schedule without the daily review limit set)