Bug#783929: [mnemosyne-proj-devel] SM-2 bug with patch
But rule 6 says not to apply rule 5 for failed cards: If the quality response was lower than 3 then start repetitions for the item from the beginning without changing the E-Factor It doesn't say that. It says that rule 6 applies only to cards with grades 0 and 1. It doesn't and shouldn't say anything about the applicability of rule 5 in rule 6! As I see it, it does, because rule 6 says not to change the easiness for failed cards, and changing the easiness is what rule 5 is all about. Ergo, rule 6 says to ignore rule 5 in this case. But I think we can at least agree that this description is worded awkwardly and confusingly, especially if it contradicts code snippets! But if that is ok, because the precise algorithm is not important for what you want to do with this data, then there is no reason not to fix the scheduler now. I agree. BTW, I'm not claiming I'm religiously following SM2, as I've also made a few other tweaks with respect to early and late reviews. The most crucial difference (which is probably a lot more significant than the issue you discuss) is that I add some randomness in the intervals, and it is precisely this noise that should allow us to explore different variants on the scheduling algorithm. My gut feeling is that if you look at the statistics comparing e.g. the average grade on the 'correct' SM2 interval, and those 1, 2, 3, ..., 10 days shorter or longer, you won't really see a big difference... Cheers, Peter -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#783929: [mnemosyne-proj-devel] SM-2 bug with patch
Done. Peter On 3 May 2015 19:32:54 CEST, Astrid S. de Wijn astrid-spamme...@syonax.net wrote: On Sun, May 03, 2015 at 06:30:03PM +0200, Peter Bienstman wrote: But if that is ok, because the precise algorithm is not important for what you want to do with this data, then there is no reason not to fix the scheduler now. I agree. Excellent! I look forward to the update. Oops, I misread your mail as there is no reason to fix the scheduler now :-) Anyway, I'd be happy to tweak the scheduler, but I'd like to back it up with statistical data analysis first, especially since this behavior has been in place for almost 10 years, and people are used to it. People are also used to programs doing what it says in the documentation. If you aren't going to fix the scheduler, then at least mention this difference with SM2 in the docs. It is a pretty major deviation, and one sentence is easy to write. Or, better yet, you could create an option. Best regards, Astrid -- You received this message because you are subscribed to the Google Groups mnemosyne-proj-devel group. To unsubscribe from this group and stop receiving emails from it, send an email to mnemosyne-proj-devel+unsubscr...@googlegroups.com. To post to this group, send email to mnemosyne-proj-de...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/mnemosyne-proj-devel/20150503173254.GJ26143%40sliepen.org. For more options, visit https://groups.google.com/d/optout.
Bug#783929: [mnemosyne-proj-devel] SM-2 bug with patch
Hi, Thanks for looking into this! However, when I wrote this code many years ago, I had a feeling the original SM2 algorithm could use a few tweaks, especially when dealing with early / late review, and the fact that there should be a small difference between grades 0 and 1. So, I'm afraid the scheduler is working as intended as far as I'm concerned. Of course, if you really want to run an 'orthodox' version of SM2, you can easily write a plugin which e.g. implements the changes you suggest. Cheers, Peter PS: Mnemosyne's own bug tracker is here: https://bugs.launchpad.net/mnemosyne-proj -Original Message- From: mnemosyne-proj-de...@googlegroups.com [mailto:mnemosyne- proj-de...@googlegroups.com] On Behalf Of ast...@syonax.net Sent: 02 May 2015 12:39 To: mnemosyne-proj-de...@googlegroups.com Subject: [mnemosyne-proj-devel] SM-2 bug with patch Dear Mnemosyne developers, Thanks for mnemosyne! It's very handy and I have been happily using it for about 2 weeks now. However, I have found a strange bug. The E-factor is not updated in the way the SM-2 algorithm describes that it should be, especially if the uesr has a lapse. I use Debian GNU/Linux testing, which still has 2.2.1, so I filed a bug report in the Debian bug tracking system yesterday. I included a patch. I've just checked and the latest 2.3.3 tarball for linux is also affected, so I thought I should post the bug report on this list as well. You can find all the details about the bug and the patch on the Debian bug tracking system (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=783929). Best regards, Astrid -- You received this message because you are subscribed to the Google Groups mnemosyne-proj-devel group. To unsubscribe from this group and stop receiving emails from it, send an email to mnemosyne-proj-devel+unsubscr...@googlegroups.com. To post to this group, send email to mnemosyne-proj- de...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/mnemosyne-proj-devel/a7faa672-b6cc- 43f0-a3d0-2606701d5043%40googlegroups.com. For more options, visit https://groups.google.com/d/optout. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#783929: [mnemosyne-proj-devel] SM-2 bug with patch
this bug, there is in fact no difference at all between grades 0 and 1. There is no difference in terms of easiness, but cards with grade 0 show up more often in the queue. Worse, cards that the user reviews precisely on time, but fails to recall (grade 0 or 1) end up with higher easiness scores than cards that the user can recall, but barely or with serious effort (grades 2 and 3). Are you really sure that is what you intended? It is a pretty major deviation from the original SM2 algorithm. If you fail to recall the card (grade 0 or 1), the easiness stays the same, but the interval gets reset to zero. A similar thing happens in the original SM2. See rule 6 in http://www.supermemo.com/english/ol/sm2.htm Cheers, Peter -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#783929: [mnemosyne-proj-devel] SM-2 bug with patch
It is rule 5 of the SM2 algorithm that is not being executed at all for cards graded 0 or 1 But rule 6 says not to apply rule 5 for failed cards: If the quality response was lower than 3 then start repetitions for the item from the beginning without changing the E-Factor I guess the reasoning behind this was that after a lot a repetitions, increasing the difficulty as well as resetting the interval was considered too big of a penalty. I agree in the corner case you mention (immediately failure), this seams suboptimal, but remember that this is not an exact science, and the idea is that after many repetitions and corrections by the user, the intervals and easiness factors converge to something which is roughly OK. I'm hesitant to change the scheduler now after so many years without detailed statistical analysis to back up any change. The data is there in the collected learning logs, but analyzing it has not yet made it to the top of my list. Anyway, thanks for your detailed feedback! Peter -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org