@kateg It toggles showing cards that were due in the past. I’ll make Anki hide it if you have no overdue cards in the next update.
@Nebulosa thanks for the report, fix coming in the next update.
@nneffs thanks for the continued feedback. Regarding the 12 month/all history selection - it controls how much of the review history is fetched from the database. For users who’ve been using Anki for years, there can be a significant speed difference between the two. It does not alter how many cards are fetched, so graphs that are derived from card data and not the review history do not change, and they do not need ‘all history’ selected in order to show longer periods.
I’ve added a bit more of a top margin to sections and moved some of the graphs around - what do people think? I personally found the old order to be a little more aesthetically pleasing, but it does make sense to put the most frequently referenced stuff near the top. Are there people here who reference the card counts frequently?
Thanks for adding the number of days studied to the stats page.
One more problem: it says I’ve studied N+1 days out of N days (instead of N days out of N days). It either added one day to the numerator or subtracted one day from the denominator.
I find I am missing the pie chart…can we bring it back? Or have a user option to choose pie or bar. I read the linked article on why pie charts are deceptive to the eyes because it is difficult to get an accurate sense of things. However, since this isn’t a scientific endeavor that requires precision, I think the pie chart is more motivational and adequately represents the data.
I miss the average ease statistic. Can you just put that under the ease graph?
And what i find kinda weird: On mobile at the very top is the name of the deck of which i am seeing the statistics. But when i click on “collection”, it still shows the name of that prior deck at the top. Would make more sense if it just shows “collection” aswell.
It would be nice if the new pie chart would highlight in bold the selected type when hovering (with the mouse) the chart and the type name as the old line chart:
@dae could you please make some sort of line that indicates “we’re only showing this much of the calendar not because you haven’t used Anki before that day, but because you set the stats to 1 year only”
I noticed that the ease graph is centered around 250 and not the actual average ease. Is that intended? I think it would make more sense if it was centered around the average ease.
Regarding the topic of the new graphs, specially the introduction of calendar (i.e. sort of heatmap).
I want to share an article I recently read that was shared in a discord group about the Abuse of Metrics in Learning. It’s a short read, I recomend to read specially the headmap section and meditate about it.
Article summary
Progress in learning is rewarding, but progress is generally slow and the reward of learning is delivered irregularly. Metrics can be used to reify progress by accelerating the feedback loop between learning and reward. People abuse metrics by substituting intrinsically meaningful goals with arbitrary metrics. The obsession with metrics may have been accelerated by social media and a general lack of meaningful goals amongst learners using SRS. I want a culture shift in the kinds of content large SRS communities value.
As the author says, users may share their stats online feeding the other side of the coin:
How often is it that people make SRS progress posts describing how learning with a Spaced Repetition System contributed to them writing a great piece of research, starting a business or writing a book? Hardly ever. Instead we are bombarded with posts that are essentially along the lines of “Look at how great I am for persisting through an endless grind controlling a flashcard slot machine”.
So, one of the caveats of the calendar graph is that may encourage users to review daily religiously for the shake of stat padding i.e. without real learn drive or motivation to do so, potentially playing a long-term role in burnout or lack of enjoyment using Anki. Anki subreddit is getting new users quickier than ever and this kind of stat-sharing is on the rise as well.
My input:
Note that my post is for vanilla Anki. There are add-ons that the whole purpose is sharing stats and even competing and comparing stats-wise against other Anki users; as that would fall outside the scope of this topic.
That is one of the reasons I try to reveal few stats of my own collection (rounded down count of cards and reviews) and usually I share some details in private but never graphs or numbers. If you saw some screenshots of graphs of mine when beta-testing 2.1.28 I cropped the pictures to show what was needed to point out and some others cases where the full graph was necesary, they were from a tiny deck or filtered search. On the other hand, I don’t remember exactly where but in this discourse, I said that I missed 5 days but I ommited the actual streak that I may have or other details.
I think the article is a bit unkind. Rare is the creature that can toil away at something without a sense of progress, and it’s natural for people to be proud of the effort they’ve put in. Some do get a bit caught up in the game side of things, but assuming they’re just a slave to the system is silly - if they did not have long term learning goals, they would not be using an SRS.
I’d like to expose a personal example for how I understand this “slavery”:
Back in the day, perhaps 15 years ago when I had access to to internet at home for the first time, I discovered MMORPG games. In this genre, you eventually do quests for the shake of gaining experience points and it’s a real grind with an absolute lack of enjoyment, hence you can strictly say you are more “working to get the daily quest reward” rather than beig laisure time. Nowadays I know people in the same loop with Pokemon Go for instance.
An analogy example for stat padding would be NBA player Russell Westbrook. He usually prioritizes stat padding to get a triple-double, even if the has to dimish team play. In that sense, you could argue that his personal goal is not necesarily winning the game (memorizing in SRS) but "being top #1 in triple double stats (stat-focused SRS goals). His end goal in reality is a mean to achive the real end which is winning the game.
Hi @dae
I find I am missing in the new stats graphics how many hours studied in past 30 days, 1 year, and lifetime. Currently, I can only see the number of reviews. It is more meaningful to know how many hours I have studied than how many cards I looked at. It is especially meaningful because I am studying a foreign language and would like to know how many hours I have spent learning the language.
Percentage correct. I use an expanded stats add-on to see this for young, mature, and total cards, as well as today’s, the week’s, and month’s (or year’s) average. I’d rather see what % I got right, not wrong.
I use the heatmap add-on to watch my attendance. It lists both streaks and percentage. I mostly watch percentage.
The forecast graph. It’s not essential, but I’ve always kept an eye on the ratio of mature to young cards. I don’t see the point of removing this. Sometimes you get a spike and can see what it’s made of.
The history graph. I use the various trend lines to track the effects of any changes in my habits. I recently backed off some overly aggressive studying this summer, and I’ve been holding off on new cards until things become consistent again, and I’ve been using these to know when my reviews have become “normalized” again. If I make a change it’s easier to see the effects when lines that were parallel start converging or vice versa. A stacked bar graph can look to erratic to show the patterns I’m used to looking at. Or to see the ratio of relearns vs mature cards, or to see if I’m overweight in young card reviews compared to another time.
Same graph, I don’t have new cards every day, but there are always mature, young, and relearns. Squeezing new cards underneath relearns makes reading the relearns inconsistent if you don’t have new cards every day. Especially without the trend lines.
The pie chart. I don’t really care if it’s a pie or a bar. The blue of the new cards “feels” more positive to me than the ominous black it used to be. I think it makes more sense to have the young portion between the new and mature, and “eat away” at the new portion, like it was. Maybe I just need to get used to it, but if feels easier to read the old pie. Mainly here, I watch my total mature cards, and my young cards. To avoid pushing myself, I’m only adding new cards if young cards is below 1000. My progress will be determined by how quickly cards either become mature or are leeched.
My main indicator of progress is net mature cards, from this add-on. The other graphs it provides helps me see what the source of progress or stagnation might be from.
Time spent reviewing and # of cards reviewed being combined in the way they are means I can’t save both in a pdf periodically (I preferred images I could share online, honestly). Although I don’t watch time as closely, you’re still giving me less information to work with than before.
The vague grey trend shadows feel a lot less substantial than the lines I’m used to.
The gradient across the bars just makes it kind of hard to read at the light end.
These new graphs are huge. Previously I could view up to four at once, but now it’s just almost three. I don’t even really care about half of them, like added, ease, or time of day, although the ease visualization is interesting.
The orange of the cram reviews really blends in with the relearns, tho I’d actually rather not even count cramming on that graph.
Graphs don’t remember the time period I’ve selected for them. This might mean a lot of redundant clicking to get the stats I want every day.
In general, I’ve wanted to track my accuracy over time, so a graph that does that would be nice. Also my card types. A stacked bar graph that shows the transition from new to young to mature/leeches over time would be really neat and satisfying. Numbers, not percentages, so it grows as I add cards. I understand you’re not looking for feature requests right now, but there ya go. Maybe that’s a solution to the pie chart.
For all the complaining I do want to thank you. I’ve been using Anki for about ten years now. This was a jarring change. Also I’m sorry if this seems strangely belated, but I guess the changes just hit my distro (I use Arch btw) in the last week or two.
I love the new “Review” graph with the different colorings for Mature, Young, Learning, Relearning, and Early categories. It would be great if you could apply that to the “Future Due” graph so we can see the distributions of future Mature, Young, and Relearning cards, not just the totals, straight away.
I’d like the Card Ease graph to be expanded down to 0% too (My cards go down to 50% and I can’t see them on the graph at all because it stops at 100%). The better option would be to allow the user to determine the X axis numbers.
I’m not sure how useful dividing up the future due graph would be, as it’s only showing the current state, not predicting what will be in the future.
Cards are supposed to have a minimum ease of 130% - an ease of 50% would mean a Good answer got a shorter delay than the previous one, which makes no sense.
