This one is tricky :) I should have commented on this RE.
First, you are right - the $ is redundant with anchoredPattern. I forgot to remove it from the committed diff.
Answering your question about %11 and %12, here is the trick. Those are placeholders replaced at runtime with whatever evaluates out of the QString::arg(...) function invocations that follow in order. If fact, the first placeholder is not %110 but %11. I started numbering from two-digit numer, here 11, not from 1, on purpose. The following 0 in the RE is literal digit that the RE expects after decimal point represented by %11. According to documentation (https://doc.qt.io/qt-5/qstring.html#arg-12), QString::arg(...) interprets placeholder strings %1, %2, ..., %99. Therefore, I had to number placeholders from two-digit numer to avoid interpreting %10 as a whole.
Now let me explain what the RE is intended to do. It says something like: Match any string that starts with 0, then has a localized decimal point, and then optionally a few more zeros but no more than our precision minus one. What for? Because I do not what 0., 0.0, 0.00 and so on to be successfully converted to double and immediately rejected by the result >= bottom() test. If that would happen, the user would be unable to input a value such as 0.01. Thus, strings like 0.000 (assume precision=4) ought to be validated as QValidator::Intermediate, which means incomplete but can easly become QValidator::Acceptable if the the next entered digit is non-zero.
When it comes to capturing groups. Well, I could do that. Instead of matching a sequence of zero after decimal point, I could capture any decimal part and reuse it later but still I would need to do another RE matching to check if the captured decimal part is a sequence of zeros to allow 0.0... as Intermediate. I am not sure if it is worth it. In the end of the day, Regular Expressions are not the easiest to read and understand (as might be proven on our example :) ). And also I doubt REs are more efficient than the classic snippet found in lines 413-419.

Thank you for the explanation. Completely valid. Agreed, a comment would help. I did not know about the trick to start with a two digit placeholder index.

One problem though: QRegularExpression::anchoredText only appeared in Qt 5.12 and we still support older versions (e.g. 5.9), so we should use ^ and $ as part of the RE.