With the arrival of 4th Edition, the ground has shifted under our feet. New rules mean a new meta, but more importantly for this blog, they mean new data structures.
In my previous analyses, such as Unit Cost-Effectiveness: Army vs Army Comparisons and the Clash of Kings 2025 Update, I ranked armies based on a “standard” composition of Regiments and Hordes. But 4th Edition has blown that structure up. We don’t build armies by just grabbing three regiments and a horde anymore; we build them using the new categorical building blocks: Core, Auxiliary, Specialist, and Support.
To keep up, I’ve rewritten my simulation scripts to reflect how we actually build lists now. The goal remains the same: to find which armies offer the most raw mathematical efficiency.
The New Methodology: Adapting to the 4th Edition Structure
The old “Top 3 Regiments + Top 2 Hordes” approach is dead. It doesn’t reflect the new unlocking mechanics or the restrictions on Specialists and Support.
I updated the Python script to pull the most cost-effective units from each of the new 4th Edition categories. My goal was to simulate a “Power Build”—if you were trying to maximize the raw efficiency of your list, what would you take?
Selection Criteria
Instead of arbitrary slots, the script now selects the top performers from each category to create a composite “Army Score”:
- Core: Top 2 Units
- Auxiliary: Top 1 Unit
- Specialist: Top 2 Units
- Support: Top 2 Units
The “Best-in-Slot” Weighting (The Spam Factor)
I also tweaked how I calculate the final score. In a competitive list, you rarely take a “diverse spread” of units for the sake of variety. You identify the most efficient hammer or the most efficient chaff, and you take as many as the points limit allows.
To reflect this behavior, I applied a heuristic weighting system:
- The #1 Unit in a category gets a 1.5x multiplier.
- The #2 Unit (where applicable) gets a 1.0x multiplier.
This heavily favors armies that have at least one absolute superstar in a category (The “Spam Factor”), rather than armies that have a deep bench of “pretty good” units. If your best Core unit is mathematically broken, your score goes up significantly—because realistically, you’re going to base your entire list around it.
The Rankings
Based on 2.5 million combat simulations and the new weighting system, here is how the factions stack up in raw efficiency.
KoW 4E Unit Cost-Effectiveness
Analysis & Disclaimers (Read Before You Rage)
Before you run to the store to buy a Kingdoms of Men army based on that 16.8 score, let’s talk about the giant, math-shaped elephant in the room.
1. The Kingdoms of Men Outlier
Kingdoms of Men sitting at 16.4 is… suspicious. This is likely an artifact of the Elo per Point metric. The model loves dirt-cheap units that perform even marginally well. If KoM has a 70-point Core unit that punches slightly above its weight, the math amplifies that efficiency because the denominator (points) is so low. It doesn’t mean KoM will win every tournament; it means they pay very little for their stats. We’ve already seen several lists spamming halberdiers and dogs of war to create some truly ridiculous US builds.
2. The “Spherical Cow” Problem
As always, this model simulates units fighting in a vacuum. It loves high Defense and high Crushing Strength (Empire of Dust, Undead). It hates synergy.
- Nightstalkers (2.6): They are at the bottom because their strength comes from Mindthirst and stealth—none of which show up in a raw “stat-check” simulation.
- Northern Alliance (7.4): Similar issue. If your army relies on aura stacking or complex interactions, this model will undervalue you every time. Similarly, armies with a variety of plausible builds instead of one spammable build score lower from a pure stat perspective.
3. Heuristic Weighting is an Art, Not a Science
The 1.5x / 1.0x weighting split is admittedly arbitrary. I didn’t run a regression against 4th Edition tournament wins (because they haven’t happened yet). I chose these weights because they felt right for how competitive players build lists: you identify the broken unit, and you lean on it. But this is a heuristic, not a law of physics.
4. Data Parsing in a New Edition
We are still scrubbing the data for the new edition. There are bound to be parsing errors or keyword interactions that the simulator doesn’t perfectly capture yet. Treat this as a “Beta” release of the rankings. As Elliot fixes the various errors (and I can load them into my model), we’ll see shifts in units and thus armies.
Conclusion
The data suggests that Empire of Dust, Goblins, and Undead remain “stat-check kings” going into the new edition. If you want to play an army where the math is inherently in your favor—high defense, high efficiency, and strong grind—these are safe bets.
If you play Nightstalkers, Forces of Nature, or Northern Alliance, don’t panic. Your low ranking just means you have to work harder for your wins. You can’t rely on raw stats to carry you; you have to use movement, terrain, and synergy to outplay the spreadsheet.
