So I’ve noticed in my Android app (haven’t tried iOS yet) that every time I switch out of the app to do anything else on my phone and when I switch back to LingQ the app moves back a page or sometimes HALF the lesson. So when I swipe back to the correct page I was on then I notice when I click to finish the lesson it has odd numbers for the amount of times I read the lesson. It’s like: 1.11x , 1.5x, 1.8x, etc. because it’s recounting those pages the app automatically switched me back to and then I had to swipe past them to the actual page I was on but I didn’t re-read any of these pages. And I can’t reset this to just 1.0x in the app. I believe I can on the web version from what I recall. So then when I’m done reading for the day I have to manually go onto the LingQ website, then add up and fix my words read statistic. This is a huge pain. Please fix this bug or whatever it is because the Word Read stat is a very important statistic which I focus my daily/weekly/monthly studying goals around. With this issue being SO bad sometimes it’s going to make people think they’ve read FARRR more than they really have. For example many people are proud to announce reading read 1 million words when maybe now with this current problem they will have only read 600K - 700K and come to realize their stats don’t really line up with their lower comprehension level than they achieved with reaching 1 million with the previous version of LingQ. Thanks
Edit:
I should add I do appreciate how a lesson doesn’t need to be completed to get those words added to your stats for every swiped page. So the fix to this issue should be to make it not possible to have more than 1.0x read before you click the finish lesson button. Only when you reopen the lesson again for a re-read does the word count start going up
Nope, I wished there was this on the web version but there isn’t. If I read faster (for example with this ultrareading we were discussing) the lesson goes 0.6x or 0.8x and there is no button to make it just 1.0x.
In the Spanish challenge there is a user with 1+ million coins, maybe he didn’t even realize that the words count is off the chart. And this is not the only case.
I simply have given up on the words count statistics for now. Fortunately is not a parameter that I need to be precise with. But otherwise I can understand your difficulty.
It is the same issue on the web version but I simply reduce the same total amount by 20% by the end of a reading session for the day and be at peace mentally and get on with my life. It is not a perfect solution but at least my reading count is not inflated. So much clicking and less language learning on LingQ 5.0 :).
Yes, on the web you can. First you need to click the minus button to reset the counter to zero. After that, clicking the plus button will increment the counter by 1.
Update: This doesn’t work as expected, don’t do this. Sorry!
The reading statistics work perfectly fine for me on iOS btw.,
frustratingly this is not the case on the web version. Today I started to read a little again in Traditional Chinese and, after increasing the font size, my read count jumped to 18x for that lesson… Correcting the number manually has unfortunately not reduced the number of coins I “earned”. At the time of writing, I have more than 6000 for today.
The effect this behavior has on the challenges is indeed annoying. In consequence, I have to make sure I’m not reading Chinese Simplified on the web for the next 2 months, because I entered a challenge in this language. Thus far I have successfully eschewed the web version there, hopefully I won’t slip up
I usually keep a running total of how off the reading counter gets and fix it at the end of the day, but I also do more and more reading completely outside LingQ and then just estimate by my reading speed * time read.
The auto counter is getting better, but I wish it simply let me reduce the count to 0 and manually set to 1 at the end of a lesson.
This actually doesn’t work like it should. If you go down to zero and then increment by one this isn’t actually reflected correctly in your reading stats. It just adds an extra read-through.
Say you read a lesson 1x but the counter says 1.2x. So you try to fix it by pressing minus until it says zero and then + so it says 1x. If you check your words read stats for the day it will have counted that lesson 2.2x. So basically by doing this you just make your stats way more inaccurate than they already were.
I wrote a comment about this bug a month or two ago but it never got fixed.
I do hope all this word count stuff gets fixed at some point because these small inaccuracies on each lesson over a long period of time are going to make everyone’s reading stats way off to the point of it being pointless to even track them.
I just tested this out too! Yep… Still bugged… So minusing from the read counts of lessons does absolutely nothing to your actual words read count. It only changes the superficial read counts of individual lessons. So in other words, it’s useless. I wonder how much of my wordcount is inflated… Hmm… 10%? I guess it really depends how much reading you do on your computer and how much on your phone. I noticed ages ago with the double counting on the Android app every time you go away and return to the app. I was just minusing 1.0x every second lesson or so, because you can’t, annoyingly, minus 0.3x. But luckily I wasn’t doing the minus then plus counts to correct this, like some others above, who were triple counting! Maybe I need to do some listening and reading in sentence mode for a while to try and correct my words read stats, as reading in sentence mode doesn’t record your words read (bug). Though, I know some people like sentence mode, but personally I don’t find it great (i.e. useless and never use it), cause I can’t look up definitions of words I’ve forgotten… These statistics bugs are pretty bad. I mean, some of us mainly use LingQ over other softwares because of the statistics. It’s one of LingQ’s main benefits. These stat bugs really need to be fixed.
While we’re on the topic of statistics bugs, LingQ does not record the listening time of the YouTube video you listen to, but rather the length of the audio file. This is a problem when the YouTube video is long and it’s automatically broken up into multiple lessons. Eg. A video I imported is 15:27 long and it’s broken up into two lessons, the first lesson having an audio file of 11:52. If I watch the entire YouTube video through LingQ with the first lesson open, my listening statistics only record 11:52, not the actual length of the YouTube video 15:27.
Youtube videos imported using the LingQ extension are a curious case. I don’t watch the embedded YouTube videos, I just use the imported text and audio. I have some recent imports that behave as “stand-alone” lessons, that is the each audio file has exactly the length of the lesson.
Most LingQ 4 imports unfortunately present themselves in such a format that includes the complete audio file in the first lesson. The subsequent lessons are trimmed at the start, i.e. they start at the nominal beginning but are “open-ended” as well. I recently revisited one of the latter type. This one specifically: Login - LingQ This is lesson 1, it has a runtime of 1:00:17, despite the video having been split into five lessons. Another annoyance is that when revisiting the lesson, the audio slider sits at the point you left off, typically around the 15 minute mark and requires manual resetting. Of course all of this is old news and unchanged from LingQ 4. The importer has changed multiple times since then.
The real problem is indeed the listening statistic, after reading and listening each one of the five lessons in sequence, the system recorded something like three hours of listening instead of just one. This was on the iPad app. I didn’t report the bug earlier, sorry.
Addendum:
In fact, just opening the lesson on the web has resulted in 13 minutes of listening time and 35 words of reading according to LingQ, I just pressed space to see the audio duration, which, speaking of the devil, happens to be bugged as well and shows 00:17 instead of 1:00:17. (might be related to zoom set to 125%) Imgur: The magic of the Internet
Ugh… Indeed buggy. I just realised that on the desktop version, when you add words read and hours listened, you update the total statistic (as opposed to on Android, where you ADD listening hours and reading time). I just removed 10% of my words read since the 5.0 update (most reading is done using the browser), as per the bug @jjacks303 mentioned. The graph now looks particularly funny.
Sorry about the late reply here, everyone, I missed this thread, I apologize for that.
I am looking into this and trying to figure out how to reproduce all the issues you mentioned on this thread. We need to have this fixed and word read/listened counters improved.
Thanks for following up zoran. I have no idea how to reproduce the 0.9x situation and yesterday I even finished a lesson and it said I read 0.5x. Maybe it happens if I don’t click on a single word on a page. No idea. I am looking at every page and not skipping any pages so it doesn’t make sense. Anyway that’s a rare issue. But most of the time I’m getting 1.1x to 1.2x words read per lesson and that’s easy to reproduce by simply going back one page (on web or android) or on the android app when you leave the app and then reopen it LingQ goes back a few pages and recounts all those pages as words read.
I already identified some problems. We will fix the problem on Android and it will be possible to reset the counters back to 0 there.
On the web, we will also make changes and once the read/listened counters are manually reduced on the complete lesson page, stats will be reduced too to follow that number.
We will also look into that going back pages issue.
I already identified some problems. We will fix the problem on Android and it will be possible to reset the counters back to 0 there.
On the web, we will also make changes and once the read/listened counters are manually reduced on the complete lesson page, stats will be reduced too to follow that number.
Also, our Android developer just confirmed that the app not reopening on the exact same page issue is fixed for the new version that will be pushed to everyone soon.
While you’re checking out this stuff, maybe look into somehow moderating coins for words read for people in challenges who suddenly enter in 2 million words read in a single day but only have 3 new LingQs and 191 new Known words in the same week. If you want to see an example, look at the Spanish Hardcore Challenge right now for May and April. The number 1 guy is ranked 76th in LingQs and 16th in Known Words but whatever he did to add 2 million words to his reading score makes him skip by everyone else and rank #1 by over 100,000 coins. It’s a silly thing to care about, I know, but then what’s the point of the challenges if they can be messed with this easily and demotivate others from trying harder?
We have been trying different implementations. Currently the Android app counts the page as read using a timer without requiring the user to page. There is a bug however. The timer starts again even if that page was counted already if the user puts the app in the background and comes back. There is also another bug which put the user in the wrong page when they came back. That is fixed for the new version coming in the next few days to everyone.
Do you think only counting the page after swiping AND completing the lesson (or pressing on View Lesson Stats button at the last page after already completing) is a better solution?
We implemented it this way so that it would count without requiring the user to page or complete the lesson. We can also adjust the interval that you can increase or decrease the read times to 0.1x and go all the way to zero, and we can also force 1.0x at least when first completing the lesson so that 0.8 or 0.9x for example doesn’t show up when you complete a lesson. Let us know what you think.
I’m not sure the best implementation of it, but probably don’t require the ‘complete lesson’. I’ve seen in the forums that some people don’t use the ‘complete lesson’ button. This is because they use LingQ for extensive reading (that is, they don’t lingQ every blue word, only the most important ones, leaving some still blue) and there is no way to turn off ‘move blue to known’, when you click finish lesson (there should really be an option to turn this off).
Don’t force 1.0x with audio at the end of a lesson. Personally, I often skip introductions and music, so I sometimes finish lessons with 0.8x or 0.9x listened, but that’s accurate. Also, probably don’t force it with reading too, because many people reread the text several times before clicking ‘complete lesson’.
Is there some way to stop the timer starting again, when the user puts the app in the background? Such as, temporarily record what the last page counted was. If the last page counted was page X, don’t count page X again.
It is surprising that the Android app shows these problems, the iOS app works perfectly in this regard, for example I can read as fast as I want, when the page is turned - it counts as read. I have never see any 0.x values, while this is the norm for me on the web. Maybe the iOS implementation can serve as example?
I understand there is a desire to acknowledge partial reads in the statistics, but, and I am repeating myself, an option for power users to disable any timers and have the system just count the complete lesson as read, once the finish lesson button is clicked, would be appreciated. This might also free up valuable developer time, as they wouldn’t have to spend so much time diddling with increasingly complex heuristics that try to guess when something has been read - I believe many users would be happy to bear the onus themselves.
Of course others like nfera actually like the incremental counter, so to reconcile all the different ways LingQ is used it might be best to, give users more control over the reading statistic, it is after all subjective. So, having the controls on the finish lesson page actually work would already go a long way. That includes the coins, e.g. decreasing the times read counter from 1 to 0 should also remove coins, this is particularly valuable when participating in challenges, as false values distort the results (see Stewart’s post). Naturally there needs to be a sanity check, the user should not have the ability to enter absurdly high values.
