Some more explanation improvements

### Description

A recent Explanation change I done, (#16229) has led me to more explanation improvements.

I am thinking of looking at the following:
1. [ ] Testing gap in testExplainFailingOptionalClauses. It only asserts on the failingOptionals list, not the main subs list. Adding a `subs.clear()` into BooleanWeight doesn't fail any tests.
- https://github.com/apache/lucene/pull/16509
2. [ ] CheckHits.checkNoMatchExplanations is shallow. It only checks the root explanation's isMatch() == false and never looks at its sub-details. Nested non-match sub-explanations are never structurally verified.
3. [x] Improve the KNN explain message. When testing #16229, we saw some unclear output, see below.
- PR merged: https://github.com/apache/lucene/pull/16271
```
0.0 = no match on optional clause (track(knn=DocAndScoreQuery[1404635,...][0.7655482,...],0.7655482))
0.0 = not in top 1 docs
```

4. [ ] TestComplexExplanations has a long-standing `// :TODO: we really need more crazy complex cases.`

Choose one unchecked item first. Start with the named testExplainFailingOptionalClauses, CheckHits.checkNoMatchExplanations, or TestComplexExplanations entry point and inspect BooleanWeight where relevant; done means the tests cover the listed explanation structure or add the requested complex cases without weakening existing assertions.

Written by the indexing model from the issue text.