When a card is initially entered, Anki requires reviews after just 1 minute and then 10 minutes. After those reviews the interval between reviews rises substantially, to 1 day. The interval expansion rate after that may vary a little*, but for my cards the typical expansion rate is by a factor of about 2.4 for each successful review. That means that successful reviews will raise the interval to 2.4 days, then to 2.4 * 2.4 = 6.76 days, and so on. On average, I get about 1 in 12 cards wrong, so by the 12th card we're up to about 2.4 9 = 2,642 days between reviews. Note that we raise to the 9 th power rather than the 12 th power, because it's not until the third repetition of a card that the interval reaches 1 day. If you sum those intervals all up, it suggests the typical time between failed reviews is about 12 years. Note, however, that I haven't been using Anki for nearly that long, and this estimate may be over-optimistic. We can get a lower bound on the time between failures by observing that my mean interval between card reviews is already 1.2 years. To achieve an interval of 1.2 years requires about 0.9 years of successful prior reviews, so on average my cards involve at least 2.1 years between failures. However, the real number may be much higher, since there's no reason to assume my next review on most of those cards is going to fail. So let's say that a conservative estimate is a mean time between failures of between 4 and 7 years. If we assume the mean time between failures is 4 years, then over 20 years that means 5 failures, and reviewing 5 failures * 10 reviews per period = 50 times, for a total of 50 * 8 seconds = 400 seconds, or about 7 minutes. If instead we assume the mean time between failures is 7 years, then over 20 years that means roughly 3 failures, and reviewing 3 failures * 11 reviews per period = 33 times, for a total of 33 * 8 seconds ≈ 260 seconds, or about 4 minutes. Note that in Anki's model a failure resets the review interval back to 10 minutes, then to 1 day, 2.4 days, and so on. In practice, that seems much too conservative. After one or two failures with a card I usually catch on, and it would be better if Anki wasn't so draconian in resetting the review schedule. A better review schedule would reduce the total study time, and I wouldn't be surprised if a typical commitment of ˜2 minutes was possible.