There are two aspects in which similarity enters the infringement analysis:

1. As evidence of copying, and
2. When determining whether what was copied was enough to be infringement

Let's take it for granted that you've based your item off of Leo's baton. So, your question is about similarity at the second stage, after copying has been established.

All US courts use "*substantial similarity*" as the test for this stage. (This can be confusing, because some courts use the same phrase for the amount of similarity required in the first stage, to establish that copying happened).

Substantial similarity can arise even when taking only a small percentage of the original, especially if you've taken the "heart" of the work. ([Harper & Row v. Nation Enterprises 471 U.S. 539 (1985)](https://supreme.justia.com/cases/federal/us/471/539/case.html) - where 300-400 words of a 7500 word manuscript were taken).

Various circuits have developed their own specific tests that juries should use to assess substantial similarity. In the 9th Circuit, two cases that highlight their approach to similarity between works of art are:

* [Sid & Marty Krofft Television Prods. v. McDonald’s Corp., 562 F.2d 1157 (9th Cir.1977)](http://law.justia.com/cases/federal/appellate-courts/F2/562/1157/293262/)
* [Pasillas v. McDonald’s Corp., 927 F.2d 440 (9th Cir.1991)](https://casetext.com/case/pasillas-v-mcdonalds-corp)

Krofft established a two-part test that separates idea and expression, such that copyright in a piece of art would not preclude people from taking that idea, but would protect the particular expression. You are not prohibited from making a baton, even if your inspiration was Leo's baton. However, if you take too much of the particular expression of Leo's baton, that crosses into substantial similarity. Whether you've taken too much is a question for the jury. "There is no special standard of similarity required in the case of 'things'."

Pasillas uses the Krofft test to judge whether two masks are substantially similar. Given that both parties conceded that their works both shared the same idea, the court could skip to "whether the masks' expressions of that idea are substantially similar". But, it incorporated elements of another line of cases that further precluded a finding of infringement when expression necessarily follows from idea.

> [w]hen idea and expression coincide, there will be protection against nothing other than identical copying of the work

> no substantial similarity may be found under the intrinsic test where analytic dissection demonstrates that all similarities in expression arise from the use of common ideas

> elements of expression that necessarily follow from the idea receive no copyright protection

> the court concluded that the similarities between the toy dinosaurs — their postures, body designs, and cuddly softness — all derived from the idea of a stuffed toy dinosaur, and accordingly found no substantial similarity of expression

> all of these similarities derive from the common idea of a mask depicting the man in the moon. Pasillas simply cannot rely on these standard elements to satisfy the intrinsic test of substantial similarity of expression

When creating a baton, there are certain elements of the expression that are inseparable from or are standard elements of the idea. It needs to be narrow. It is standard to have a handle. The handle is often a different color.

In summary:

* Just because you saw the baton, and decided to make a baton based on what you saw, that isn't necessarily infringement.
* If you take too much of the particular expression of Leo's baton, that will be infringement.
* The answer to that question is case-by-case, and left to the jury. "Obviously, no principle can be stated as to when an imitator has gone beyond copying the idea, and has borrowed its expression. Decisions must therefore inevitably be ad hoc." - Judge Learned Hand