I would set desired retention to 99% and forget that I have a life.
Guess I am effed then
Maybe itâs better to max out higher order learning and to reduce the remaining cards further by chunking, mind-mapping etc. ⌠But I guess you already did so âŚ
So, greater Anki review efficiency would give you the opportunity to add even more cards 
.
(And DR(90%) or DR(85%) should be more efficient than DR(99%)âŚ)
Letâs say your workload with DR(99%) will be 120% compared to the workload with DR(85%). (I think thatâs a conservative estimate ⌠but it still proves the point.)
You could take your weak spots and add slightly different cards that also cover those weaker topics.
Taking the graph from Expertium as an estimate for average retention (AR), one could argue, that the overall remembered cards with DR(85%) at the same overall review workload would be greater compared to DR(99%)
- DR(99%) â AR(99.5%)
- DR(85%) â AR(92%)
Your total number of remembered cards at the same overall workload would be:
- DR(99%) â 15000 * .995 = 14925
- DR(85%) â 15000 * 1.2 * .92 = 16560 (> 14925 !)
(Itâs not about the exact numbers of course ⌠but about the Idea to be able to add more cards while having the same workload.)
Of course you also have to take into account the effort to create additional cards ⌠but I think that a great deal regarding learning happens during this card creation.
However, I donât know about Anatomy or other Med School topics ⌠maybe itâs just lotâs of material where you cannot work through it in peace ^^.
Thatâs very similar to what âCompute optimal retentionâ is doing. It tries to find retention that maximizes the sum of all retrievabilities (R).
For example, if you have 10 cards and their Rs are 90%, then the sum is 10â˘0.9=9.0. If you lower retention, you can squeeze more cards into the same timeframe. Say, you have 12 cards with R=80%. 12â˘0.8=9.6. The sum is greater, therefore learning 12 cards at 80% desired retention is better than 10 cards at 90% desired retention.
But again, it only goes as high as 95%.
Okeeeeey. Now I get it. But why does it say experimental though. What about it could be further changed
And if it is about me getting as much cards (pieces of information) as efficiently as possible inside my poor puny miniscule smoothbrain of mine, this clears a lot of things for me.
For example, I estimate my deck for the locomotive system to be of 8000 cards. I need to get this done by the end of February-beginning of March. (lets say about 30 days, with 180 min per day study time)
Based on this, my DR should be 0.93
But what confuses me is twofold
-Why is it that when I turn it into 40 days, it decreases to 90%
But when I increase it further to 80 days, it becomes 95%
I dont understand whether increasing or decreasing the days increased DR or not and why.
-Secondly, why does increasing Minutes per day increases the DR. I also dont understand the reasoning behind this.
Of course you also have to take into account the effort to create additional cards ⌠but I think that a great deal regarding learning happens during this card creation.
I think I have already reached a high level enough of atomic cards, where I cannot break it down any further. I split my sentences in a prose into cloze cards, such that not too much context is shown but also just enough. I feel like I cannot do much more.
So what do you reckon, how should I tackle the remaining 5-7% estimated cards that I will be forgetting because of this.
However, I donât know about Anatomy or other Med School topics ⌠maybe itâs just lotâs of material where you cannot work through it in peace ^^.
It is hell manifest on Earth.
With more study time, you will run out of the new cards of the deck quickly. And the spare time will be waste if you only do the review with a low retention. So the optimal retention will increase.
1 Like
Right now it assumes that you study âby the clockâ and if your allocated time of studying per day is 30 minutes but you spent 20 minutes, you add more cards. This is not how most people use Anki.
Oh and what about this
But what confuses me is twofold
-Why is it that when I turn it into 40 days, it decreases to 90%
But when I increase it further to 80 days, it becomes 95%
Is this a bug? @Expertium
Iâm sure youâre aware, but just to reiterate:
Since itâs an oral exam youâre preparing for, I wouldnât recommend spamming more Anki cards to gain the last 2-3%. Chances are, youâre just going to start recognizing the cards by pattern and not because of their actual content. Since youâve failed the exam twice already, I assume you have a good foundation in the theoretical part but may struggle to express it verbally.
For Anatomy exams, itâs important that you can present your knowledge in a coherent way. Itâs not essential to know every single last detail if you understand the connections and the big picture. Details are important, but you should consider spending more time articulating and presenting the 90% you already know rather than trying to squeeze the last few drops out of the deck.
Just my two cents from a fellow med school student.
Good luck on your exam. Youâre going to do great, Iâm sure of that!
4 Likes
Details are important, but you should consider spending more time articulating and presenting the 90% you already know rather than trying to squeeze the last few drops out of the deck.
I do struggle in this regard quite a lot, more so because I am studying Medicine in a foreign language (German). Nonetheless, I do struggle greatly with forgetfulness and after failing such a strenuous exam twice after all the effort that I have put into it, it made me doubt myself massively. I view this exam more as a threat now than a hurdle in my studies.
One particular aspect I do struggle with though is long processes or generally any lists where items have to be sequentially memorized. From the process of glycolysis to the pathway of the vagal nerve and activation of the complementary immune system.
I am sure that you, as a fellow Med student, have come across countless processes and lists. I cannot for the life of me get them inside my head.
I am now trying cloze-overlapping and flow charts combined with image loci to utilize my spatial and image memory. Have you got any tips?)
The searching method is based on simulation where the reviews are sampled from probability distribution. So the simulation results are not perfectly accurate.
1 Like
I am just going to go with the model then, since it seems I know much less than I thought I did.
Is there any chance that FSRS dynamically changes the learning or relearning steps for each card which had a history of multiple lapses in the near future? Or is this also part of the still non-existent short term memory model?
In simple terms, LMSherlock is saying that there is no exact formula for computing optimal retention with infinite precision (like Ď, for example), and it has to be estimated using a simulation.
And yes, since we donât have a model of short-tetm memory, learning and re-learning steps are unaffected by FSRS.
I have no idea about anatomy or how does your exam look like, but I donât think the retention rate is the crucial factor in your case. I cannot think of an exam that one would fail even if they gets 95% of the answers correct.
Since you passed the written exam but failed the oral one, I assume that a relevant thing here is how fast you recall the knowledge (in paper exams you can think well before writing down the answer, but in speaking you have to come up with the answer much faster). A method of training on this is changing your definition of buttons when answering cards. For example, consider using the following definitions:
- Easy: correctly answer immediately
- Good: correctly answer in 1s
- Hard: correctly answer in 3s
- Again: failed to answer in 3s, or the answer is wrong
This is probably not what FSRS or Anki is designed for, but I think its works reasonably good (I use this method to train listening in learning a language). Of course you may need to change the numbers above according to the content of cards, the requirement of the exam, etc.
Another thing to notice is that, according to my own understanding, FSRS (or space repetition in general) is not designed for memorizing everything. Instead, they are designed for memorizing a large portion (e.g. 90%) of content with relatively less effort. Probably the only way to get an ideal (~100%) grasp of some complicated subject is keep using it in everyday life.
A third note is that FSRS (and space repetition) is only a tool to remind you to review at the optimal time, it does not help memorization itself. Compared to struggling with the retention rate or other parameters, itâs usually more effective to change the method of memorizing. A simple examples is changing type of questions (cloze vs. explaining a concept from scratch). Itâs highly personal and depends on the type of content though, so (sadly) I cannot provide specific advice here.
And finally (and maybe most importantly), as already pointed out by others in the thread, exams (especially oral ones) are usually not about pure memorization. Since you already failed 2 times I assume you already have a good idea on what the questions look like. Make up some sample questions by yourself (or better, using questions of exams in the past if you have them), and try to answer them. The point is to know about what you are lacking (cannot recall quickly, does not know what to say, cards not covering exam requirements, etc.) so you can work in the right direction (which is very unlikely to be âgetting a 99% retentionâ).
2 Likes
Nah, it should be fine.
Mathematically speaking, as retention approaches 100%, interval lengths approach 0 and reviews/day approach infinity. So yeah, 100% retention is unobtainable.
1 Like
Well, I guess this means that I should just leave it at 95% and that my problem lies beyond the scope of Anki. 
Could you at least maybe give tips about memorizing long lists, pathways, and processes where everything is sequentially ordered and cannot be transposed?
I have tried everything from overlapping cloze cards, to flow charts to evoke my spatial memory and imagery as a mnemonic locus. I tried acronyms and number-letter pegs but the sheer amount of stuff to be learned makes it altogether more confusing.
I find others not having as much trouble with memorizing lists as much as and I wanted to use Anki as a tool to help me. I lose particularly a lot of time in memorizing those and almost to no avail or after a lot of valuable time lost.
I am feeling down and the stress is piling out as my exam is in two months.
I cannot think of an exam that one would fail even if they got 95% of the answers correct.
My professors are extremely - and I mean EXTREMELY - demanding. The exam lasts about 5-10 minutes and anything can be asked about in a cadaver. Near perfection is a requirement to pass the exam as there is only so much that could be asked in 5-10 minutes. Hence, my unrealistic expectations of Anki and its algorithm to achieve this high standard goal.
I guess I may not pass this exam after allâŚ
From my own experience, memorizing a sequence of things usually involves understanding the mechanism behind that, i.e. why they are arranged in such a way (by scientific rules, the design of the inventor, how itâs used in the real world, etc.). Once that is clear the whole sequence may be built on a few central steps, and other steps come up naturally by their intrinsic logic. My major is mostly related to natural science (where things are generally more logical) so it probably doesnât apply to you though.
Sorry to hear that, but that does not mean you should feel desperate now. Again idk about your exam, but assume that the exam includes 10 questions (I think thatâs quite a lot for a 5~10 min oral exam), and you have a 95% chance of answering any given question correctly, then the probability that you get 10/10 in the exam is ~60% â and another 30% chance for a 9/10 result. So I (still) think the problem is not in the retention rate, but rather how to prepare for the exam itself (âconverting a 95% retention rate to a 95% chance of correct answerâ). Try asking your schoolmates who have passed it, regarding how they prepared for the exam. Or go even further, ask the professor directly â from my own experience, many professors are willing to provide suggestions on how to prepare for the exam and/or learning the subject âin the correct wayâ (YMMV though).