Wild Tesla Energy Economy: Decay, Charge-Aware Routing, and 9-Hour Coil Durability

[rebaserHEAD]

Follow-up to the containment fix in Triad-Sector#433. That PR makes a correctly-built cage actually hold the ball; this issue is the other half, making the tesla worth caging in the first place. The goal is a tesla that runs wilder than other forks but only when you feed it hard, and that stays interesting across a 9-hour round instead of a 1-1.5h one.

The regime ladder we want, keyed off PA strength:

• Standby / Level 0: the ball starves and collapses.
• Level 1: break-even. The ball lives but generates no motes. Idle, not dead.
• Level 2: moderate output.
• Level 3 (max PA): wild, and sustainable at that output. This is the "we're cooking" state.

Right now the ball has no passive drain, coils fill on a single strike, the strike target is picked without regard to charge, and coil durability is tuned for a short round. So the whole thing is flat: no reason to push the PA, no reason to add coils, no visible feedback when the engine is saturated.

Scope (each of these is a separate knob, can be split into its own PR if it gets big):

• Passive energy decay. The ball bleeds energy over time so a running PA is required to sustain it. Pin the drain to the Level-1 feed rate so Level 1 is exactly break-even (alive, no motes), below it dies, Level 2+ nets positive.
• Mote wildness. Mote (mini-ball) spawn scales with how much energy headroom the ball has, tuned so max PA sustains a target mote count. More motes than other forks, but only at full feed. Open number: target Level-3 mote count (working default ~20).
• Charge-aware lightning routing. Coils with the biggest potential gap (most headroom) are likelier to be struck. An empty coil is 100% and top priority; a saturated coil floors out. Cleanest as a new LightningStrikeAttemptEvent hook on the strike roll, which is upstreamable on its own.
• Saturated-coil arc indicator. When a coil can't accept a full strike, show the existing arcing/charged visual. The TeslaCoilVisuals.Enabled state is currently dead (nothing sets it, no RSI state) so we can repurpose it to Charged, driven off the coil's charge-changed event.
• Coil durability for 9h rounds. Coils are balanced for an hour to ninety minutes; a coil should survive a full shift. Drop damageFromLightning accordingly. Edison staff stretch service intervals by buying and placing more coils, which is the intended sink.
• Coil power economy. chargeFromLightning (2M) currently exceeds maxCharge (1M), so one strike tops a coil out and flattens the gradient the routing above depends on. Re-proportion so a strike is a fraction of capacity (chargeFromLightning = maxCharge / F, F around 10-20), and raise maxSupply / maxCharge so a charged coil actually pushes meaningful power into the grid.
• Rare crew strikes. Players are already on the strike table at priority 1 but take zero damage (the strike applies Structural, which no-ops on Biological mobs). Add an electrocution path so a saturated engine with its rods overflowed can occasionally zap crew. Two open sub-decisions: does the crew strike chain onward or terminate, and electrocution vs a plain shock.

Explicitly cut: grounding-rod cooldown. Rods are fine as they are.

Not committing to final numbers here. The mote target and the two crew-strike sub-decisions are the open questions to settle before building.