Change starting adjustment

[justintondt]

I'm not familiar with Python. Is there a way to change the code to start with a predefined success rate (ex. 0.85) on the first answer so that a minimum amount of reviews before adjustment isn't needed?

[eshapard]

The problem is that without some reviews, we have no data about the card to use to predict a success rate. The best you could do is to use what you know of other cards in the deck and assume that new cards will be the same. If you read my blog at eshapard.github.io, you'll find two articles/addons that will help with that. One sets the initial ease factor to the average ease factor of the mature cards in the same deck. The other creates learning steps based on the initial ease factor.

Excerpts from justintondt's message of April 19, 2020 8:30 pm:

I'm not familiar with Python. Is there a way to change the code to start with a predefined success rate (ex. 0.85) on the first answer so that a minimum amount of reviews before adjustment isn't needed?

-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/eshapard/experimentalCardEaseFactor/issues/11

[justintondt]

I've read the articles and understand that without some reviews there is no data to predict success rate. However, using a predefined success rate as the first value, decreases the wild swinging of the first several values as well as prevents the rate from ever reaching 100% or 0%. The predefined number is arbitrary, but starting with the goal number makes sense to keep the values close to it. See the attached image for a side by side comparison with a random answer sequence.

[eshapard]

You seem to be suggesting assigning a default success rate to the first successful review instead of calculating an actual one. Then from there, you seem to be calculating success rate some other way, such as a moving average or something. That's a big departure from the way the addon is coded, but it could be done.

Excerpts from justintondt's message of April 20, 2020 9:12 am:

I've read the articles and understand that without some reviews there is no data to predict success rate. However, using a predefined success rate as the first value, decreases the wild swinging of the first several values as well as prevents the rate from ever reaching 100% or 0%. The predefined number is arbitrary, but starting with the goal number makes sense to keep the values close to it. See the attached image for a side by side comparison with a random answer sequence.

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/eshapard/experimentalCardEaseFactor/issues/11#issuecomment-616654662