1. Purpose
To determine whether the population of documented Civil War military and administrative figures can support a draft-and-roster game on the scale and cadence of a contemporary fantasy sports league, and to specify the positional structure that would make such a game generate productive argument rather than trivia recitation.
The design goal is not simulation fidelity. It is argument density — the number of defensible disagreements per unit of play. A good fantasy league is a machine for producing structured disputes about valuation among people who already know the material. That is the thing being built.
2. The Sizing Question, Answered First
Warner’s standard reference volumes list roughly 583 Union and 425 Confederate officers who held full general rank, before counting the several hundred additional brevet appointments, the colonels who commanded brigades in fact if not in grade, the naval officers, and the civil and administrative apparatus on both sides.
Raw headcount is therefore not the constraint. The binding constraint is documentation depth — how many figures have enough surviving record that two informed people could disagree about their value and both cite evidence. That number is realistically 300 to 450.
Standard league arithmetic:
| Format | Teams | Roster | Slots filled | Pool needed |
|---|---|---|---|---|
| Standard | 12 | 18 | 216 | ~350 |
| Deep | 14 | 20 | 280 | ~450 |
| Dynasty/keeper | 10 | 28 | 280 | ~450 |
Verdict: comfortably feasible at 10–14 teams. Twelve is the recommended size. Sixteen is possible but begins to break the premium positions. Twenty is not viable without expanding into brigade command and state-level administration, at which point the marginal picks become names nobody can evaluate and the argument dies.
The pool depth is uneven by position, however, and that unevenness is the design’s central asset. It should be engineered, not smoothed away.
3. Positional Structure
The instinct to slot everyone as “general” should be resisted. Fantasy formats work because positions have different value distributions and cannot substitute for one another. Civil War command divides naturally into functions with genuinely different talent curves.
Recommended starting lineup (12-team, 18-man roster, 11 starters):
| Slot | Starters | Viable pool | Scarcity |
|---|---|---|---|
| Army Commander | 1 | ~28 | Extreme |
| Corps Commander | 2 | ~75 | Moderate |
| Division Commander | 2 | ~180 | Low |
| Cavalry | 1 | ~35 | High |
| Artillery Chief | 1 | ~18 | Extreme |
| Engineer | 1 | ~16 | Extreme |
| Logistics / Ordnance / Commissary | 1 | ~15 | Extreme |
| Naval | 1 | ~30 | High |
| Staff (chief of staff, adjutant, intelligence, topographical, medical) | 1 | ~45 | Moderate |
| Bench | 7 | — | — |
Notes on the individual slots:
Army Commander is the premium position and should be capped at one starter with a hard limit of two rostered. The genuine candidate list runs to roughly twenty-eight names across both sides and the whole war. In a twelve-team league that means twenty-four of twenty-eight are rostered and the last team is starting Nathaniel Banks. This is correct. Every fantasy format needs a position where the drop from tier one to tier four is a cliff.
Artillery Chief is the sleeper design element. The pool is genuinely thin — Henry Hunt, E.P. Alexander, Stephen D. Lee, Charles Wainwright, Freeman McGilvery, John Pelham, and then a rapid descent. The twelfth team drafts William Pendleton, and everyone at the table understands what has happened. Thin positions with a well-known bottom are the best-loved slots in any fantasy format.
Logistics / Ordnance / Commissary is the slot that separates serious players from casual ones and is the strongest argument for including it. Montgomery Meigs and Josiah Gorgas were arguably more consequential to their respective war efforts than most corps commanders, and Lucius Northrop was more damaging to the Confederacy than several Union generals. A format that rewards recognizing this is doing pedagogical work for free.
Engineer carries the same virtue in miniature, with the added wrinkle that Herman Haupt’s railroad management may deserve carving out as its own category if the league grows.
Staff functions as the flex position and absorbs the intelligence, medical, and topographical figures — George Sharpe, Jedediah Hotchkiss, Jonathan Letterman, John Rawlins, Walter Taylor, Charles Marshall. It also absorbs Samuel Cooper, the highest-ranking officer in the Confederate army by seniority and the single greatest late-round joke pick available in either war.
A civil/political slot is deliberately omitted from the starting lineup. Lincoln, Davis, Stanton, and Seward operate on a different scale and their inclusion collapses the valuation exercise — Lincoln is worth more than any three generals and there is nothing to argue about. Recommend they be excluded entirely, or confined to an optional single “Cabinet” slot in a variant format where their scoring is walled off from military scoring.
4. Draft Format
Auction, not snake. This is the load-bearing recommendation of the prospectus.
A snake draft produces a ranking. An auction produces prices, and prices are the interesting output. With a fixed budget, every participant must state numerically what they think Lee is worth relative to Grant, and relative to Thomas, and relative to two corps commanders plus an artillery chief. The disagreements become quantified and comparable across the table.
The auction also solves a problem specific to this material. Reputational premiums in Civil War historiography are large, unevenly distributed, and mostly inherited from postwar advocacy rather than from performance. An auction makes the premium visible as a number. When Longstreet clears at a third of Jackson’s price, or when Hood costs more than his 1864 record can possibly justify, the format has produced a measurement of something historians usually argue about qualitatively. Running the auction, then scoring on record-based metrics, yields a residual — market price minus earned value — that is the closest thing to an empirical handle on the Lost Cause reputational economy that a parlor game is ever going to generate.
5. Scoring
Three viable architectures, in ascending order of construction cost.
(a) Record-based static scoring. Each figure is scored once from historical performance against defined metrics: engagements as commanding officer, force ratios faced, casualties inflicted against casualties sustained, objectives taken and held, promotion trajectory, and duration of retained command. Rosters are compared by sum. Cheap to build; the arguments migrate to metric design, which is where you want them. Downside: the season is decided at the draft.
(b) Head-to-head scenario matchups. Each week two rosters are placed against one another in a defined situation — a river crossing under pressure, a siege, a raid against a rail line, a defensive stand with inferior numbers — and each participant argues how their roster handles it. Adjudicated by rubric plus league vote. This is the format best matched to a discussion group and the one recommended for a first season. Construction cost is a scenario bank, which is the same work as writing wargame scenarios.
(c) Chronological season. The league runs 1861 through 1865 in monthly or quarterly turns. Figures become unavailable when they died or were removed — Albert Sidney Johnston in April 1862, Jackson and Reynolds in 1863, Stuart and Sedgwick and McPherson and Cleburne in 1864. This maps precisely onto injury risk in conventional fantasy and creates genuine draft tension between peak value and availability. Whether Jackson is worth a top-three price for twenty-four months of service is exactly the kind of question this format exists to generate. Highest construction cost; highest payoff.
Recommend building (a) as the valuation substrate, running (b) as the weekly engine, and holding (c) as a second-season variant.
6. Known Design Problems
Scale drift. A brigade in 1861 and a brigade in 1864 are not the same object, and command of one in the first year is not comparable to command of the other in the last. Any record-based metric must normalize by year or the early-war figures will be systematically mispriced.
Sample size. Some figures held independent command for a single battle. Their record-based scores are noise. A minimum-engagement threshold is required, with a separate designation for high-variance small-sample figures who become the format’s lottery tickets.
Replacement level. Fantasy value is value over replacement, and defining replacement level for a Civil War corps commander is not a game-design question — it is the underlying historical question, unresolved. This should be treated as a feature and made an explicit league debate rather than settled by fiat.
The both-sides question. Rosters should draw from both armies. A format that segregates them is a reenactment, not a fantasy league, and forecloses the comparisons that make the exercise worth running.
Handling the genuinely repugnant. Some drafted figures were slave traders, war criminals by the standards of their own moment, or founders of terrorist organizations. A game format can launder this by treating the men as trading cards. The design answer is not to exclude them, which distorts the historical picture, but to keep scoring strictly keyed to military function and to make the reputational-premium mechanic explicit in the rules. A format whose stated purpose includes measuring how reputations were manufactured is doing the opposite of laundering. This should be written into the prospectus front matter rather than left implicit, because it will come up and it is better answered in advance.
7. Phased Development
Phase 1 — Pool construction. Assemble and tier the figure database by slot. Target 350 entries with a one-paragraph evaluation and a preliminary tier assignment each. This is the bulk of the work and is reusable across every format variant.
Phase 2 — Valuation calibration. Run a mock auction with four to six participants to establish price ranges and identify positions where the tiering is wrong. Expect the artillery and logistics slots to require the most adjustment.
Phase 3 — Scenario bank. Twelve to sixteen matchup scenarios with adjudication rubrics.
Phase 4 — First season. Twelve teams, auction draft, head-to-head scenario play, single-elimination playoff.
Phase 5 — Instrumentation. Record auction prices against record-based scores and publish the residuals. This is the output with value beyond the league itself.
8. Open Questions
- Should brevet rank count toward positional eligibility, or does the Union brevet system inflate the pool past usefulness?
- Are naval figures scoreable on a common scale with land commanders, or do they require a parallel scoring track?
- Is a keeper format viable given that the underlying figures never change, or does the absence of rookie classes make dynasty play static?
- Does the format extend cleanly to other conflicts with comparably documented command structures, and if so, is the reputational-premium measurement portable?
9. Recommendation
Proceed. The pool supports twelve teams without strain, the functional divisions of Civil War command produce a positional structure with naturally uneven scarcity, and the auction format converts participant disagreement into comparable numbers. The unexpected byproduct — a quantified measure of the gap between a figure’s market reputation and his defensible record — is worth more than the game.
