Document that chained `.run_if(a).run_if(b)` don't short-circuit, in contrast to `run_if(a.and(b))`

## How can Bevy's documentation be improved?

Fun fact:

This panics
```rust
fn noop() {}
app.add_systems(Update, noop.run_if(|| false).run_if(|| panic!()));
```

this does not:
```rust
fn noop() {}
app.add_systems(Update, noop.run_if((|| false).and(|| panic!())));
```

Would you have guessed so? :D The docs should probably be on `run_if` and point users to `and`.

The reason for this is that the `run_if` calls need to update their ticks, so that things like `Changed` behave as expected. Does `and` have that problem? IDK
One person says "It looks like the return value of and is also a System, so it enters the ECS as a single system, with one set of ticks", the other "I think each sub system of and will store its own ticks, so the second system will have an old value if it gets skipped."

Locate the existing documentation for `run_if` and `and`, then compare the two examples in the issue and investigate the stated system-tick behavior. Done means the `run_if` documentation explains that chained calls do not short-circuit like `run_if(a.and(b))`, points readers to `and`, and accurately addresses the tick caveat.

Written by the indexing model from the issue text.