Subformula

In ludics, formulas are represented as loci ξ and subformulas by subloci ξ*i. Immediate subformulahood respects focusing: so P⊥, Q⊥, and R⊥ are all immediate subformulas of ((P⊥ ⊕ Q⊥) ⊗ R⊥), since ⊕ and ⊗ are both positive connectives.

One way to look at this is that the above formula could have been written as However, if you look at the leaves of the left-focusing step, they are exactly the same - two possibilities, one of which has "|- Δ1,P" and "|- Δ1,R" as leaves, and another which has "|- Δ1,Q" and "|- Δ1,R" as leaves. The idea behind loci is to observe that because the behavior of the focused rules is exactly the same, it is counter-productive to view all 12 formulas as 12 different things - the use of loci has the result that all 12 must be treated identically.
 * ((Q⊥ ⊕ P⊥) ⊗ R⊥)
 * (R⊥ ⊗ (P⊥ ⊕ Q⊥))
 * ((P⊥ ⊗ R⊥) ⊕ (R⊥ ⊗ Q⊥))
 * etc... I believe there are 12 ways in all.

You can see how this turns out for a left-focus on ((P⊥ ⊕ Q⊥) ⊗ R⊥) below: