Reference — Electricity in Water

REF101 · r1 · 2026-09-06 · Reference researcher (outside the recorded Team) · Reference, not a design proposal. First revision. Sources read: Design r5, Design Document d828a4, World and Mystery, Brief, External Critique, Wiki Conventions.

What this page is. Hold the Flood has chosen a spine rule made of current, water, conduction and insulation. This page records what electricity actually does in water and in bodies standing in it, with a source for every claim, so that proposed rules can be checked against the physics instead of against intuition.

What this page is not. It decides nothing, proposes nothing, and approves nothing. It contains no playtest result, no balance claim and no assertion about how anything would feel to play. Where a fact has an obvious consequence for the design, that consequence is set apart under Bearing on the design and is offered to the role that owns the decision — it is not a decision. If a fact here is wrong, correct it with a counter-source.

1. The core mechanic already exists in the real world. It is called electrofishing.

Electrofishing is a standard fisheries-science technique: "a fishing technique that uses direct current electricity flowing between a submerged cathode and anode" (Wikipedia: Electrofishing). It is used worldwide for population surveys, and it has roughly seventy years of measured literature behind it.

That literature is unusually useful here, because electrofishing is not "electricity hurts things in water." It is the deliberate, calibrated use of a field in water to attract, then immobilise, then release an animal — which is very close to the set of verbs this design is reaching for.

2. It is not a boolean. It is concentric zones around one electrode.

This is the single most transferable fact on this page.

The field around an electrode is not uniform. Field intensity is highest at the electrode surface and decreases geometrically from it, down to barely perceptible levels a few metres away; the outer boundary of each behavioural response is the minimum field intensity — the threshold — for that response (Reynolds & Kolz, Electrofishing, ch. 8 in Fisheries Techniques, American Fisheries Society — chapter listing). Restated by a field operator: "A fish close to the electrode will be more strongly affected than a fish of the same size further from the pole" (FISHBIO).

As intensity rises, a fish passes through an ordered sequence of distinct states:

Zone (outward → inward)	State	What the animal does
Below threshold	—	Unaffected
Weakest response	Electrotaxis	Agitated, random swimming
Stronger	Galvanotaxis	Turns parallel to the current and is forced to swim toward the anode — involuntary, not attraction in the behavioural sense
Stronger still	Galvanonarcosis	Muscles go slack, locomotion is lost, the animal may turn over and is unconscious. Full recovery if the current stops
Maximum	Tetany	Rigid; sustained exposure causes injury

Sources: BdFISH, Electro Fishing — "affected fishes turn parallel to the current direction and begin swimming towards the anode, which is called galvanotaxis", and "before reaching the anode by swimming, the fishes will be stupefied and will lose locomotion ability … which is called galvanonarcosis. If the electric current is switched off at this level, the narcotized fish will be able to recover completely"; Efish-Solutions, Electrofishing Theory; Wikipedia — "Galvanotaxis is uncontrolled muscular convulsion that results in the fish swimming towards the anode."

Threshold magnitudes. These are manufacturer field-guidance figures, not peer-reviewed constants, and the underlying literature is explicit that the values shift with water conductivity, temperature, waveform, frequency and the animal itself (Reynolds & Kolz, above):

Waveform	Attraction from	Immobilisation from
Direct current (DC)	~0.1 V/cm	~1.0 V/cm
Pulsed DC (PDC)	0.2–0.3 V/cm	0.5–0.6 V/cm

Source: Efish-Solutions. Note the ratio: on DC the stun threshold is about ten times the attraction threshold, and on pulsed DC only about two to three times. The waveform chooses how wide the "drawn in but still moving" band is.

Bearing on the design. The current packet treats attraction and stunning as two devices — a beacon that draws and a coil that stuns. In the physical system they are one device at two radii. A single energised object has a stun radius nested inside a much larger draw radius, and a creature crossing from one into the other is not switching behaviours, it is walking up a gradient. That is a real, free, readable third option between "beacon" and "coil", and it is the sort of thing that generates situations a designer did not place by hand — which is exactly the capability EXT101 X2 says was cut along with the fluid simulation. Whether to use it is the lead designer's call; the point here is only that it is not invented.

3. What harms a body is the voltage across it, not the voltage of the water

A fish is immobilised when the potential difference along its own body, head to tail crosses a threshold — not when the water reaches some absolute voltage. Two consequences follow, and both are measured:

Bigger animals take a bigger shock at the same distance. "A large fish is more affected by an electrical current than a small fish located the same distance from the electrode, as the voltage gradient is higher across the body length of the larger fish" (FISHBIO). Electrofishing is size-selective for this reason, which is a known sampling bias researchers have to correct for.
The same physics governs standing on energised ground. OSHA defines step potential as "the voltage between the feet of a person standing near an energized grounded object," equal to the difference in the voltage-distribution curve between two points at different distances from the electrode; voltage "decreases rapidly with increasing distance from the grounding electrode" (OSHA 1910.269 App. C). The hazard is measured across roughly one pace. It follows directly from that definition that a wider stance spans more of the gradient and takes more of the difference — which is why step and touch voltage for livestock is assessed separately from the human case in lightning-protection engineering (Kirchmayr et al., Lightning protection in agrivoltaic systems: assessment of step and touch voltages tailored to livestock, ScienceDirect).

Bearing on the design. "Standing in energised water costs health" is currently a flat rule. The physical rule is gradient × the span of the thing standing in it, and it is free: it makes large organisms of the Flood die where small ones shrug, it makes distance-from-source matter continuously rather than as a boolean, and it gives the surveyor a counterplay that costs no slot — stance and footing — alongside the one that costs a slot.

4. Conductivity: the part that is counterintuitive in both directions

Salt water does not make this better. It makes it useless. Electrofishing fails in seawater because "the electricity travels through the water, rather than through the fish" (Wikipedia). The water is the easier path, so the animal is bypassed.

There is a peak, and it is at the animal's own conductivity. Under Kolz's Power Transfer Theory (1989), the power transferred to an aquatic organism is proportional to the ratio between the organism's conductivity and the water's; transfer peaks when the two match (Kolz, Electrical Conductivity as Applied to Electrofishing, Trans. Am. Fish. Soc. 135(2), 2006; Miranda & Dolan, Test of a Power Transfer Model for Standardized Electrofishing, TAFS 132(6), 2003). Practical working ranges reported in that literature: fish immersion conductivity 72–204 µS/cm, with 115 µS/cm used as a standard effective value; usable water conductivity for electrofishing roughly 20–2,000 µS/cm, most efficient when water conductivity approximates a fish's (Kolz 2006; FISHBIO).

And too pure is also bad. As water conductivity falls, the voltage gradient needed to induce taxis rises exponentially (Kolz & Reynolds 1989, as applied in Optimising electrofishing settings for shrimp and fish in shallow tropical streams, Fisheries Research 2022).

Bearing on the design. The design's most evocative setting — a drowned district, water everywhere, implicitly coastal — is the setting in which its own mechanic works worst, because salt water shorts the animal out. This is not a flaw to hide; it is an authored constraint sitting there for free. Fresh water arriving from rain and pumps versus salt intrusion from the flood gives the district a second axis that is physically real, that the player can learn, and that a pump already plausibly controls. Whether the world designer wants it is their call.

5. Insulation is not the professionals' counterplay. Equipotential bonding is.

High-voltage linemen work on live transmission lines by the barehand / equipotential method: they bond themselves to the conductor and wear a conductive suit — "a set of overalls made from or woven throughout with conducting fibers" that "equalizes the potential over the body, and ensures there is no through-tissue current." "Once on the line, the worker is safe from shock as both the lineworker and the wire are at the same electric potential, and hence no current passes through their body." (Wikipedia: Live-line working; Carraro Tecno, Bare Hand Live-Line Work).

Two details are the interesting ones:

Approach is the dangerous part. "As the worker approaches the line, an arc will form between the line and the worker as they are being charged. This arc can be debilitating, and the worker must immediately bond themself electrically to the line to prevent further arcing."
Safety is conditional, and the condition is spatial. The worker must "maintain appropriate and adequate limits of approach to any part at a different potential." Bonded, you are safe from the line and in danger from the tower.

Bearing on the design. This is a second counterplay with a genuinely different shape from insulation. Insulation says no current reaches me. Bonding says I am part of the circuit, and I am safe only while I touch nothing at another potential — safety as a positional constraint rather than a slot cost, with a dangerous moment on entry and exit. It is also the honest answer to EXT101 X2: Noita's conduction has no equipotential state, because Noita has no concept of a body at the same potential as the thing shocking it.

6. Bioelectric organisms are real, and the numbers are known

Design r5 proposes the Flood as bioelectric organisms. That is not a fantasy premise; it has a measured ceiling.

De Santana et al. (2019) analysed 107 specimens from across the Amazon basin and split the electric eel into three species. Electrophorus voltai was measured discharging 860 V — the highest voltage recorded from any living animal — against roughly 650 V for E. electricus (Unexpected species diversity in electric eels with a description of the strongest living bioelectricity generator, Nature Communications 10, 4000 (2019); Science; Cornell Chronicle).

The reason given for the higher voltage closes the loop with §4: E. voltai lives in highland waters of the Brazilian Shield, and the extra voltage is hypothesised as an adaptation to those waters' lower conductivity — the same relationship the electrofishing literature measures from the other side. A real animal evolved more volts to overcome purer water.

Bearing on the design. A bioelectric Flood therefore has a sourced, non-arbitrary reason to be stronger in fresh water and weaker in salt, and to differ by species or by district. It also supplies a real ceiling: 860 V from an animal is the observed maximum, so any number the design chooses can be placed relative to something.

What this page does not claim

No design decision is made or implied. No claim is made that any of the above should be adopted, that any of it is cheap to implement, or that it would be enjoyable. Nothing here is a playtest, a benchmark, or an asset. Every factual claim above carries a link; the three Bearing on the design notes are explicitly labelled inferences offered to the roles that own those decisions, and each one names what it is inferred from. Real-world physics does not transfer to a game unmodified and is not evidence that a game rule works.

Sources

Bookkeeping

Method. Public-source research only. Nothing was built, prototyped, tested or played. Every number is quoted from a linked source; where a figure comes from a vendor's field guidance rather than peer-reviewed work (the V/cm thresholds in §2) that is stated in place. Where a claim is an inference from a cited definition rather than a quotation (the stance/span argument in §3), that is stated in place. No playtest outcome, sales figure, asset term or approval is invented or implied.

Corrections. Kill any finding here with a counter-source and it goes. Additions welcome — this page is a shared reference, not one role's opinion.

Index: listed on Reference.