I’ve noticed, that lately the Flatten functionality was added to FSRS Helper: If the number of future due cards exceeds the limit, the cards are postponed.
I have a hard time understanding the advantage of this solution compared to just setting a daily reviews limit - which is a basic Anki functionality.
On a second thought, it seems that the Flatten feature has a real advantage.
The flatten feature will preferably shift the high stability cards to the future. In contrast, the review limits in Anki don’t consider the stability (unless you use ascending intervals or relative overdueness as the review sort order).
So, if the review limits in Anki “shift” the low stability cards to the future, the impact on the retention might be greater.
Using the built-in features, you can minimize the impact on retention by choosing ascending intervals or relative overdueness as the review sort order. But, in that case, you won’t be able to sort your reviews randomly.
It would be a nice addition if done in review limit too, replacing the current gather-order dependent method.
I believe Jarrett showcased something similar on GitHub once, but I haven’t looked at it personally. If it shifts due dates simply on the basis of stability as you say, what does it do with existing backlogs? I would assume something like “relative overdueness” but does it?
I think both try to do similar things but in different ways. I don’t know about how load balancer exactly works in Helper add-on but it does respect fuzz. Flatten probably doesn’t. It’s not your fault that you confused the two, people will get confused by similar sounding things.
That’s why it was great load balancer in Anki didn’t include any toggles. No confusion with review limit!