Are we talking about language learning here now, or video games?
5 Likes
Ideally theyâre the same.
1 Like
You can look at it as a Memrise speedrun, but itâs an important topic.
When learning a language youâll reach a point where you know the basics, just need to learn thousands of vocabulary items. The faster you can do, the better for you. How fast you can do? Depends on you, and depends on the Memrise algorithm. @Casper_duo is a power user who cares about efficiency - he tries to optimize his part. Memrise should care about the efficiency of all users, and their goal should be make more and more âspeedrunâ possible.
I know redesigning badges is important, but increasing overall efficiency must be more important. If an average user can learn +3 word per session during the same time spent on Memrise, that might be a huge competitive advantage.
If I were Memrise team, Iâm sure Iâd do constant tweaking of the teaching algorithm with A/B testing on different groups of users. I think the review intervals could be optimized per user, or even per word.
I donât have any goals like reaching Overlord in xx months or planting YY words per day. BUT: I need a Memrise who has these goals in the background. Who says âNah every hour you spent here you can learn 12.5 words, letâs see if I can make you learn 15/hour.â
2 Likes
But maybe it is this way for a reason? I mean, after Iâve learned something through Memrise, it sticks. I too is a sort of power user, but in the sense that I communicate with native speakers in my target language every single day. Also remember, your learning isnât limited to Memrise (or shouldnât be).
If you always answer correctly while planting, you receive (45+20)*6 points per word. The 20 points is a bonus per answer when you donât get any wrong. If you get one wrong, the bonus is 10 points per answer, and if you miss more than one, there is no bonus.
Day 1: Plant 100 words for 39,000 points.
Day 2: Plant 100 words for 39,000 points. Water 100 words (planted on Day 1) for 134+20 points per word, earning 15,400 points. Total: 54,400 points
Day 3: Plant 100 words for 39,000 points. Water 100 words (planted on Day 2) for 154 points per word for 15,400 points. Water 100 words (planted on Day 1) for 150+20 words per word, earning 17,000 points. Total: 71,400 points.
Days 4-8: (like Day 3)
Day 9: Plant 100 words for 39,000 points. Water 100 words (planted on Day 8) for 154 points per word for 15,400 points. Water 100 words (planted on Day 7) for 170 words per word, earning 17,000 points. Water 100 words (planted on Day 1) for 170 words per word, earning 17,000 points. Total: 88,400 points.
Days 10-20: (like Day 9)
Day 21: Plant 100 words for 39,000 points. Water 100 words (planted on Day 20) for 154 points per word for 15,400 points. Water 100 words (planted on Day 19) for 170 words per word, earning 17,000 points. Water 100 words (planted on Day 13) for 170 words per word, earning 17,000 points. Water 100 words (planted on Day 1) for 150+20 words per word, earning 17,000 points. Total: 105,400 points.
After 24 more days, the total per days is 122,400 per day, and 48 days after that, the total is 139,400 per day.
If you get a word wrong in a watering session, itâs worth 45 in the session where you get it wrong, and the next time you review it, itâs worth 54 points. After that, itâs 68, 92, 112, 134, and finally 150 again.
If you miss the word in the first couple weeks after planting it, the extra watering sessions do not make up for the points lost.
Itâs missing out on the bonus of 20 points per word for getting all answers correct that makes the most difference in the long term.
1 Like
Iâm tempted to say stop worrying about points and just learn/review. With that, the points will follow. However, itâs also possible that itâs the points that get you to keep learning so who am I second guess that.
1 Like
Using points to set goals and measure progress, I was able to motivate myself to learn thousands of words in multiple languages, and I had fun along the way. If I were only motivated by points, I would get bored rather quickly.
Of course not. Learning words is part of a bigger picture that includes listening, speaking, writing, and reading. No one claimed you can learn to speak a language using only MemriseâŠ
Of course. Iâm not saying âgameifyingâ isnât good, As long as you donât âloose sightâ and become more obsessed with numbers than actual language comprehension, it is a good motivator. I know all about that. Always have to complete my 15 min a day goal in order to not break the streak.
1 Like
The first week has ended. Unfortunately, despite my best efforts, I wasnât able to beat the chart.
Did every single one of my reviews + more than 300 words today, that is about 4000 questions, but again was let down by the number of review items which was too little too late.
You just canât compete without it.
These guys are heavy hitters. I think the romzez guy did like 500,000 points today.
Not even the first 3 places after all that work. 
Thatâs for me. Fair and square. At least I know now how they do it.
Iâm also more than half a million points behind on my weekly goal, but not a bad result. I can work with that.
I had a good run with Spanish today.
A pace of 130K points/hour and a flawless score. (Donât skip your reviews!!)
But this lucrative points source was dried up as quickly as it showed up.
Since that is about 4 seconds per question you can quite probably reach better results.
Iâve also noticed a significant drop in the difficult words count from a few hundred to about 100.
Maybe now it will actually fulfill its function and show real difficult words and not just phony ones. Iâll have to check it.
@Kaspian that was a bit more detailed response than I expected haha, but thanks for the effort. Rough calculations are more than sufficient.
You count the points earned from learning with the reviews, but we are only interested in the review part so you can simplify it quite a bit.
I need approximately 100,000x2 just from reviews. And to get that you need a consistent flow of more than 1200 review items per day. It will take a while for me to reach it, as you can see from my forecast:
I just have 2 days on the horizon with more than 1000 words and I need that consistently every day.
I completly agree. The app and the web (without scripts) are very inefficient. They can do a lot to improve the time spent by their users.
It is not +3 it is more like 50% (more?) improvement. Really significant. I kinda wasted a lot of time first with the app even though it is better than the default web version.
Since I will probably close tomorrow, or the day after that, my first 10,000 words, and also finish the official Swedish course, I would say that my focus is pretty sharp, but thanks, I acknowledge your point. Iâm just having a bit of fun as to not make it too monotonous and keeping me in check. Overload is not the goal but a far a way dream (I will not make it but nonetheless will give it my best) that will push you to strain yourself to achieve an unimaginable amount (100,000 words in a year?) in a ridicules amount of time span. At least you can say you tried, right? They gamify it so why not treat it as a game? Iâm just going with the flow. In a game, you often need to adapt yourself and improve your technique to pass a difficult level. The same thing applies here. You just look for ways to do it better and faster for the same amount of work and hope to rip the end results and âwinâ, whatever that means. If you will look at it on the long run you will see that every second count and accumulates. More points represent more work so strive to get points. Simple as that. Itâs the same thing.
Btw, by which word count did some of you noticed a significant improvement in your ability to watch movies unassisted? Is there such a number? 10K? âŠ20K?
2 Likes
I think there are so many factors for movies , I think it is hard to give a clear answer. But I would say in general 10,000 words is a necessary but not at all sufficient condition. It may be one on average one of the hardest things to do in the language you are learning (especially if your goal is appreciation and enjoyment of this). But - I will say personally speaking, without subtitles (if we consider those assistance), you need to know almost all of the grammar patterns of the language and perhaps how they are morphed when used in colloquial speech. Also - your single word recognition has to be really good and I think it goes deeper than knowing one definition, but more like a few of the most common âgo-toâ definitions in your native language. Because your brain doesnât have the time to compute the most optimal translation unless you happen to have heard similar dialogue/grammar patterns before and it is automatic. Then, of course, slang / historical references / plays on words. Which I think require a ton of exposure to varied sources to begin to appreciate. But - if your goal is to use movies as a study tool and use subtitles, pausing, etc. Feel free to ignore everything I said haha.
3 Likes
âBtw, by which word count did some of you noticed a significant improvement in your ability to watch movies unassisted? Is there such a number? 10K? âŠ20K?â
I donât think the word count is the most important, rather how fast you can recall them. Iâd focus on the core 6000-8000 words, ofc the more you know the better.
- I know about 3-4000 spanish words. I can understand most of the Spanish narration on youtube videos when itâs clear speech.
- TED talks are a different level, as it is live speech. I have some problems when they talk too fast, or in argentinian accent, or thereâs noise in the background, but itâs still a good material for practicing.
- Understanding everyday âstreetâ Spanish should be the next level, when it is not a narration or speech so the speaker doesnât care about articulation. It might be the most common 1000-1500 words only, but I need to recognize them in every possible verb form and meaning and idiom. (I havenât tried any movies yet.)
I think there are three stages of learning.
1, Recognizing words: 8-10 successful reviews in a row and Memrise thinks you know the word. Review intervals get long quickly (too quickly?) if you always guess the word correctly. I try to avoid tapping tests and speed reviews. It gives me a false feeling that I know the word, but I just know how to select it - a big differenceâŠ
2, Knowing words by heart: requires 20-30 encounters at least, not just typing in or selecting on Memrise, but seeing/hearing/using the word in a sentence. Often when I hear a Spanish sentence, I recognize all words, but I still have to think about the meaning. When I need time to recall some words, I might miss the next sentence and lag behind the speaker. If I know every word by heart, this shouldnât happen.
3, Effortless listening: Even when I know all words by heart, I can follow the sentence without losing the thread, but it requires active attention. I still need to concentrate to get every grammatical feature, catching all verb endings correctly, etc. So the next level is listening passively, without paying much attention, when you can do something completely unrelated thing while still understanding everything.
I donât know how many hours I need to reach stage 3 but I guess 100-200 hours of listening needed at least on top of Memrise practice.
2 Likes
I donât know. Steve Kaufman, for example, like to say that when he reaches 30K~40K words at LingQ he has some sort of mental breakthrough and are able to understand his podcasts to a great deal of degree. I figured it should be the same here. We are both just racking up our word count. There should be some sort of a general number correlation that pushes you over the edge. You need exposure and this quantifies exposure.
I would focus on them too, but that doesnât mean that the words that are not in the top most frequency lists are any less important, as you would eventually need to learn them as well to make sense of what you hear and read. This is just prioritization. At the end, you still need to learn them all.
I donât think you can understand a whole lot just from 3000 words. That will only put you in a A1/A2 level.
Consider this. You reached the word max interval and still donât quite know it. Why does it matter?
At some point, if you donât really know it, itâs just a matter of time (be it half a year or more from today), before you forget it completely and eventually make a mistake. That is all you need 1 MISTAKE. At such a point, you will just restart the cycle from 0, and again get a lot of new exposure and it will eventually click in your head without you even noticing.
You donât need to keep track of what you really know and what you donât really know. If you donât know it you would eventually have to relearn it, so the focus on any one particular word in the grand scheme of things is utterly meaningless. It doesnât matter. You just keep progressing and trust that what needs more work will resurface naturally on its own in the future. That is why Iâm not so worried that I might be tricking the algorithm to a false sense of knowledge. I understand that I might need to practice words that I learned long ago in the past and itâs ok if itâs even hundreds, as long as my real voacb is steadily increasing with thousands of others entries that I do really know.
How high is the percentage of the vocab that does get ingrained in your brain? I donât know. But Iâm guessing itâs high enough to be worth it.
I remember that I read someone here in the forum that when he isnât absolutely sure he knows the meaning, he picks the wrong answer on purpose. That is why I donât see such necessity or concern myself with that problem.
Do you guys think that sentence courses might be better than single words courses?
I like the Memrise courses because there is a good proportion of sentences and the words that comprise them. That is a very nice method of teaching.
For example, I found this course for duolingo German sentences: Duolingo German Sentences - by lone_star - Memrise
I think it might be more worthwhile than doing the single words course.
The answers might be easier to pick in the crowd, but the exposure you get is more and also higher quality.
2 Likes
I think you are expressing a good understanding of language learning. Thereâs actually plenty of neuroscience out there about how learning, forgetting, and re-learning actually makes stronger long-term memories. But Memrise doesnât repeat the entire cycle when you make a mistake.
I personally would never select a wrong answer on purpose. I often think âI have no idea, Iâll just type something random,â and then it turns out that my typing is the correct answer. I think that is part of what has to happen to get to fluency - you stop overthinking and double-checking. Like walking, or riding a bike.
I think the Memrise courses are nicely well-rounded, but personally I try to keep a mix of courses going: a well-rounded course such as one of Memriseâs featured courses, a big frequency vocabulary course (typing), a recognition-only vocabulary course (non-typing), a few verb-conjugation drill courses (typing), a sentence course (non-typing), a small fun course with specialized vocabulary (typing). The mix would depend on the language: verb conjugations are more important in Spanish than in Japanese. And I might have 7 courses going in a language where I am pursuing fluency, but just one course in a language where travel usage is all I need.
3 Likes
I think sentence courses, ala Glossika really helps with gaining automatic understanding of certain grammar structures. I think the biggest problem with Memrise is if a sentence course is multiple choice it is easy to guess by high frequency words which diminishes listening comprehension gains, and it is a hindrance when typing if there are a lot of homophones. For example there was a really great course by Student Of Life for Mandarin sentences, but it over-uses the pronoun ta (ä»ïŒć„čïŒćźïŒwhich can mean he, she, or it. The course did not choose a single pronoun, so you have extra stuff to memorize on top of the content itself. If there are no ambiguities like that, and the sentences are authentic I think Memrise sentence courses are awesome and underutilized. But - I still think for sentences Anki is king.
For example if you want to do a sentence course on memrise which prompts on audio, then you type in your target language, you can make so many silly mistakes which give you the wrong answer (I gave one for Mandarin above). The number of alts grows combinatorially. Additionally, production of the target language is impossible on Memrise even more so due to the alts problem. So the more cognitive load you put on your brain, in terms of âextraâ stuff to memorize to get the right answer on Memrise, the more burnt out of learning you can get. Our goal at first in languages, I think, should be to be understood by a native speaker, and through iterative correction converge at proper grammar. So I think enforcing perfect grammar at the very beginning is counterproductive to learning.
Lastly - for listening comprehension and grammar recognition is where I think sentence courses can help the most. For example listening to a sentence once and dictating/typing out that whole sentence and then translating it. Sentence courses on Memrise cannot force you to both dictate and translate to get the right answer. Which is a shame, because I think parallel development of decoding the grammar as well as sound recognition is the best way to develop listening comprehension. So for these reasons - I think you should use Anki for that , also if you have the funds I highly highly recommend Glossika (though I donât know the quality of the course for the languages you want to focus on). There are tutorials on how to take the 3,000 Glossika sentences and put them into Anki flashcards so you can do SRS on them.
1 Like
I started Memrise with the official Spanish 1-7 course.
Good quality course, good selection of words and sentences, I just donât get why they limited the length of the courses. Gives 2-3 examples of grammar features, good for recognizing (reading/listening), but Iâd need 20-30 examples to use them (write/speak) confidently. Maybe the Spanish for Polyglots is a good alternative or the old A1-A2 might have more examples. (I dont have too much experience with them.)
I tried one sentence course so far:
https://www.memrise.com/course/123071/575-spanish-verbs-sentences/
Good selection of sentences, but lack of alternatives and no active topic to go with my issues. Could be a good course, just not user-friendly. Like @roflcopterlol wrote, many sentences needs memorization of extra stuff to get the right answer. Paused right now.
Current plan:
- Finish the Top 5000 words course first.
- Then try @LarBoylanâs DuoMem series, looks promiisng:
https://www.memrise.com/user/LarBoylan/courses/teaching/ - Some more sentences from the 575+ course if needed. Or some grammar parts of Polyglotâs Spanish.
- When the grammar is OK, some more words from @ian_mnâs Top Up series
If you have any suggestions, please let me know.
@VT22 Compare Swampyâs incomplete version of that sentence course: https://www.memrise.com/course/218823/spanish-listening-intermediate/
Swampy has outstanding Spanish sentence courses. Easy grammar: https://www.memrise.com/course/748004/learning-spanish-grammar-wsentences-audio/
More advanced: https://www.memrise.com/course/134555/spanish-expressions-audio/
To get the most out of these courses, donât just rush through the multiple choice, (often the multiple choice is easy). Instead, listen and repeat on every sentence, every time. Youâll be building a âit just sounds right this wayâ language intuition.
Many of Swampyâs no-typing courses are paired with a typing course by another user, but as you have mentioned, one soon runs into many alternatives with Spanish sentences.
This is contradicted by the information stated here:
I donât think it is contradictory . Some things we as humans classify as small mistakes will still be marked red by Memrise. Some omissions of letters are considered red mistakes, and sometime when I have a typo it is a yellow. Both I would classify as a âmistakeâ. The cycle (as far as I know) repeats on red mistakes, and if it is yellow you will return to the normal schedule after reviewing again. I think the first person you quoted meant red mistakes.
Iâll intentionally make some errors today and will report back in a few days.
I think the âyellowâ mistakes are also what happens when you choose the âwrongâ definition in the case of synonyms.
For example, if two words have the same definition, but are actually different words, you will get a yellow mistake.