Lanchester’s Laws: Why Force Concentration Can Beat Equal Spreading

Lanchester’s laws are simple attrition models that explain a non-intuitive strategic truth:

• in some contests, doubling force gives more than double advantage;
• in others, it only gives roughly linear advantage.

That distinction changes how you should allocate limited resources (capital, attention, engineering effort, sales pressure, etc.).

1) One-sentence intuition

When each unit can effectively target many opposing units (aimed/ranged competition), advantage scales closer to the square of force size; when engagement is mostly one-to-one and locally constrained, advantage scales closer to linear.

2) The two classic forms

Let side A have size (A(t)), side B have size (B(t)), and effectiveness coefficients (\alpha, \beta).

Linear-law style (locally constrained engagement)

[ \frac{dA}{dt} = -\beta B, \quad \frac{dB}{dt} = -\alpha A ]

(Interpretation varies by model assumptions, but practically this corresponds to environments where local matching and congestion constrain effective firepower.)

Square-law style (aimed, broad targetability)

A useful invariant form is:

[ \alpha A^2 - \beta B^2 = \text{constant} ]

So “combat power” behaves like effectiveness × force-size-squared. Small numeric edges can compound hard.

3) Why the square-law idea matters so much

If two sides have similar unit quality, under square-law conditions:

• 120 vs 100 is not “20% stronger”; it is roughly (1.2^2=1.44), i.e. much stronger effective position.
• splitting your force into separate equal fights often destroys this advantage.

This is the classic concentration principle:

• If engagement is square-law-ish, concentrate where decisive.
• If engagement is linear-ish, spreading can be less punishing.

4) The trap: people over-apply it

Lanchester is powerful, but easy to misuse.

It fails when assumptions fail

Typical assumption breaks:

1. Heterogeneous quality (skills, latency, tooling, terrain)
2. Detection asymmetry (one side can target, the other cannot)
3. Capacity bottlenecks (only N units can engage at once)
4. Reinforcements and delays (not closed attrition)
5. Morale/command collapse thresholds (non-smooth dynamics)
6. Network structure effects (not fully mixed)

Real systems are often hybrid: square-law in one phase, linear-ish in another.

5) Practical translation for builders/operators

This is where Lanchester becomes useful outside military history.

A) Product / startup competition

If user attention acts like winner-take-most and cross-network visibility is high, your market may behave more square-law-like.

Implication:

• focus effort to dominate one beachhead,
• avoid “equal weak presence” across many fronts.

B) Engineering incident response

During a broad outage, adding one engineer to five streams can underperform a concentrated strike team on the principal failure path.

Implication:

• identify the highest-leverage causal chain,
• mass effort there first, then fan out.

C) Trading / execution operations

In fragmented microstructure, some windows are square-law-ish (crowding where everyone can hit similar liquidity pools quickly).

Implication:

• avoid joining crowded fronts blindly,
• preserve optionality to shift venue/timing before toxicity compounds.

6) A quick diagnostic: is your situation linear-ish or square-ish?

Ask these five questions:

1. Can one unit influence many opponents quickly (high targetability)?
2. Does adding units increase total effective pressure superlinearly?
3. Is there low congestion in applying extra force?
4. Are interactions broadly mixed rather than segmented?
5. Does early numerical edge cascade into persistent dominance?

If many answers are “yes,” treat the regime as more square-law-like.

7) Anti-footgun checklist

Before using Lanchester logic in decision-making:

1. State assumptions explicitly (targetability, mixing, congestion)
2. Run sensitivity tests on effectiveness ratio (\alpha/\beta)
3. Test split-vs-concentrate counterfactuals
4. Check phase transitions (open/mid/late regime changes)
5. Add stop conditions for model drift

A good mental rule:

Lanchester is a regime lens, not a universal law of competition.

8) Why this remains a useful field concept

Even if you never fit differential equations, Lanchester gives a disciplined answer to a daily strategic question:

“Should we spread evenly, or concentrate decisively?”

Most teams answer by intuition. Lanchester at least forces the correct follow-up:

• What interaction regime are we in right now?

That one question prevents a lot of expensive “do everything a little” failure.

9) References (starting points)

• F. W. Lanchester (1916), Aircraft in Warfare: The Dawn of the Fourth Arm.
• J. R. N. (historical operations-research literature), foundational discussions of linear vs square attrition formulations.
• Alan Washburn, Lanchester Systems (lecture notes; clear mathematical treatment).
• Moshe Kress (2020), Lanchester Models for Irregular Warfare (Mathematics), modern extensions and caveats.

(If useful next step: build a tiny simulation notebook with phase-switching assumptions to test when concentration beats diversification under your own constraints.)