State-compression ladder · Note M14 · Part IX · generalization

How far the cut travels

Breaking one long trajectory into short blocks worked on a pendulum, and again on the Lorenz attractor. So does the cut itself generalize? Across four dynamical systems it does — but three honest edges say it is not a free, uniform law.

2026 · whitebox program generalization sweep Van der Pol · Duffing

Parts VI–VIII established the mechanism on a tame forced pendulum and stress-tested it on Lorenz. The open worry was over-fitting to two friendly systems. So we ran the same block-length sweep on two more: Van der Pol (a relaxation oscillator, dialled from mild to stiff) and Duffing (a driven cubic oscillator — strongly nonlinear but not stiff). The headline holds. The interesting part is exactly where it frays.

01Four systems, one cut

systemcharacterbest vs single shootoptimum
PendulummildinteriorT_block ~ 4
Van der Pol (μ=1)moderate9.9×T_block 4 (interior)
Duffingstrong, non-stiff51×T_block 1 (extreme)
Lorenzchaoticcategoricalshort
Van der Pol (μ=4)stiff1.5× — all poor

Wherever the surrogate integrator is adequate, short blocks beat the single long shoot — and on the most strongly nonlinear non-stiff system, Duffing, by the largest margin of all, 51×. The cut travels. Now the edges.

02Edge one — there is no universal block length

worse better VdP single-shoot level VdP best (interior) Duffing best (extreme) coarse valley — worse than single 12 48 32128 512 ← more decomposition single long shoot → block length (steps, log)
Van der Pol (μ=1), 3-seed mean Duffing (seed 0)
Free-run rollout vs block length, log–log. Both systems improve dramatically as blocks shorten, but the optimum sits in different places: Van der Pol bottoms out at an interior block length (T_block=4) and gets slightly worse at the one-step extreme; Duffing falls monotonically to its shortest block. "T_block=4 is best" was always a per-system fact, never a law.

On the pendulum and Van der Pol the best block is in the middle — short enough for a well-conditioned gradient, long enough to still compose dynamics (the two-jobs tension of Part VI). On Duffing, shorter is simply better all the way down. Same cut, different sweet spot.

03Edge two — a coarse cut can be worse than none

Look again at the Van der Pol curve where it pokes above its own single-shoot line. Splitting into just four long segments (T_block=128) trains worse than not splitting at all — and this reproduces in every one of three seeds (mean 0.606 vs 0.181 for the single shoot). A few still-long blocks, loosely stitched by a fixed continuity penalty, settle into an inconsistent solution that the fully-coupled single shoot avoids.

Decomposition is not monotone. Below some block length it wins big; just above it, a coarse split can be worse than the monolith it replaced. You have to cut past the coarse regime, not merely into it.

This is not a failure of the idea — the win at T_block=4 is ten-fold and robust — but it kills any naive "more splitting is always safer" reading. The likely lever is the coupling strength: a stronger or adaptive continuity weight for coarse splits should erase the valley, and that is its own testable claim.

04Edge three — stiffness is a different axis

Turn Van der Pol up to μ=4 and the clean story collapses: every block length is poor, the best beats the single shoot by only 1.5×, and the gradient-survival signature inverts. But this is not the decomposition failing. It is the integrator failing: a stiff relaxation oscillator is exactly where an explicit Euler surrogate is unstable, so no factorization of a broken integrator can help. Stiffness is orthogonal to the credit-path question — a separate axis that wants an implicit or adaptive solver, not a different cut.

Duffing is the control that proves this reading: it is more strongly nonlinear than mild Van der Pol, yet non-stiff — and it gives the largest win in the whole suite. Nonlinearity is not the problem; stiffness of the chosen integrator is.

05The claim, properly bounded

Short-block multiple shooting robustly beats the single long shoot across strongly nonlinear systems — provided the surrogate integrator is stable on the system, and the blocks are short enough to clear the coarse-decomposition regime. The optimal block length is system-dependent, not a constant.

That is a smaller claim than "breaking trajectories always helps," and a truer one. The mechanism generalizes; the tuning does not come for free. Both facts matter, and a note that reported only the first would be the less useful one.

06Honest edges

Duffing and stiff Van der Pol are single-seed; mild Van der Pol is three-seed and it is the one carrying the load-bearing "coarse valley" claim. All use the same explicit-Euler surrogate as the rest of the arc, which is precisely why the μ=4 result is read as an integrator limit rather than a credit result. The per-arm gradient-survival numbers are noisy (each is a property of a separately-trained field) and are not leaned on; free-running rollout is the metric throughout.


Part of the whitebox / MorphoHDL program. Generalization sweep extending the multiple-shooting arc (Parts VI–VIII). Van der Pol x' = y, y' = μ(1−x²)y − x (μ = 1 and 4); Duffing x' = y, y' = −0.3y + x − x³ + 0.2cos(1.2t); shared 2-layer vector field, explicit Euler surrogate, multiple shooting with a continuity penalty; free-run rollout the scoring metric.

AI-generated research. The experiments, code and write-ups on this site were produced by AI agents (Claude and Codex) working semi-autonomously, directed at a high level by a human. Nothing here is peer reviewed and none of it is affiliated with an academic institution. Measurements are reported as taken — including the nulls, the caveats and the corrections — but the work has had no external review.