The active force a skeletal-muscle sarcomere can produce depends on its starting length because starting length determines how effectively thick and thin filaments can form force-producing cross-bridges.

This is the active length–tension relationship.

Too little overlap gives low force

If a sarcomere is stretched so far that only a small region of thick and thin filaments overlaps, relatively few myosin heads can reach actin.

Fewer possible force-producing cross-bridges means lower active tension.

At an extreme length with no overlap, active actin–myosin force approaches zero.

Intermediate overlap permits high active force

At a range of intermediate sarcomere lengths, thick and thin filaments have extensive useful overlap without severe geometric interference. Many myosin heads can interact productively with actin, so active isometric force can be high.

Excessive shortening also reduces force

Making the sarcomere progressively shorter does not increase active force without limit.

At very short lengths, thin filaments from opposite sides can overlap one another substantially, and thick filaments can approach Z-disc constraints. These geometric effects reduce productive force generation.

The qualitative active-force curve is therefore peaked:

active force
    ^
    |        ______
    |      /        \
    |_____/          \_____
    +-----------------------> sarcomere length
      too short   too long

Passive tension has a different origin

Stretching a relaxed muscle also produces passive tension from elastic structures such as titin, a large protein connecting thick-filament regions toward the Z disc, and from connective tissues outside the sarcomere.

Passive tension generally rises as muscle is stretched beyond its slack range.

Total measured tension can therefore be thought of conceptually as

$$\boxed{F_{\rm total}=F_{\rm active}+F_{\rm passive}}.$$

The active component comes from cycling cross-bridges; the passive component comes from elastic structures resisting stretch.

The length–tension relationship is thus a structural consequence of filament overlap plus passive elasticity, not a change in action-potential amplitude.