1. Forum
  2. >
  3. Topic: Duolingo
  4. >
  5. Another (different) Question …


Another (different) Question about Skill Strength and Decay Rates

This may be a question for the Duolingo staff ...

I have been doing Duolingo for several years, and I have noticed that some of my trees decay more rapidly than I expect and other trees decay much more slowly than I expect. Of course, I realize that the decay rate depends on how many words I get wrong and how often I hover/mouse over words. Some languages (Dutch, French, and Russian) decay at about the rate I would expect, but some other languages really surprise me.

For example, I have been doing Italian since 2014. I am at level 25, have (or used to have) a reasonably high fluency rating in the high 50's, rarely miss or hover over words, and I have had no problem speaking or understanding Italian in my four visits to Italy over the past few years. But three or so Italian skills decay each day --- consistently. On the other hand, my Turkish skills decay very slowly --- more like one skill every four days. I will assure you that my Turkish skills are not nearly at the level of my Italian skills. I do my best on the Turkish lessons, but it's a difficult language for an English speaker, and I still miss a lot.

I am wondering whether the skill strength might be graded on a bit of a curve. Maybe my Turkish skill, while not perfect by a long shot, is better than the skill of an average Turkish learner? Or maybe Turkish is a shorter course, and the length of the course has an effect? Or did I do something back in the early days of the Italian course that set it on a different decay trajectory? I don't know.

Anyhow, I am curious, and I thank you in advance for any feedback or ideas.

(By the way, what I am asking about has nothing to do with the recent A/B testing for Skill Strengths. I have been noticing the varying decay rates for years.)

April 19, 2017

1 Comment


I'm wondering if maybe the way you go about practicing Italian vs. Turkish could be a factor?

If you routinely just get barely over the "regilding threshold" for Italian (which would make sense because you know it well, why do even more practice that's not really benefiting you much?) but do a lot of practice of Turkish, even if making mistakes along the way, I could see how this pattern would result.

For instance, I have spent quite a bit of time on Guarani. Because I'm very much still in the learning stage and because the course is still in an "uber beta" state with a lot of missing responses, I use an awful lot of hover hints (it's the only way to have a guess of which of several perfectly valid responses is going to be accepted for that specific question), yet (until the A/B test ending yesterday or so at least), I think only one of my golden skills had degilded since the beginning of this year. I figure it's because I go through the lessons one by one and rack up several hundred XP trying to get the content actually into my brain.

In Russian I've definitely been known to do the bare minimum to regild, after which I find that bit of work undone soon enough.

Learn a language in just 5 minutes a day. For free.