A DCI score is eight scores in a trench coat. Name a finish and see the caption roads that get a corps there, or pull one caption up or down and watch ten thousand simulated seasons re-sort the entire field — live.
26 corps modeled · same engine as the projections · updated 2026-07-28
Path to Position
Pick a corps and a finish line. The model works backward and shows the three most plausible caption-level roads there — who has to gain, who has to give ground, and how realistic each is against the way these corps are actually trending.
Start With a Question
The arguments already happening on the message boards, run through the model. Tap one to load it.
The Caption Tuner
Pick a corps, pick a caption, and move it. One caption per scenario — the whole point is to isolate what a single fix is worth.
—
The Full-Field Ripple
Every corps's title, medal, and top-six chances — before your change, and after. Green means the scenario helped them; crimson means it cost them.
#
Corps
Gold
Medal
Top 6
Baseline probabilities from 10,000 simulated seasons. Each cell shows the change your scenario produces. Distant corps barely move — that's the model, not a rounding artifact: the same simulated seasons are replayed with one number changed, so the difference is signal, not noise.
Tune the Whole Sheet
The tuner above isolates one fix. This one stacks several at once for a single corps — pull GE up, the brass up, the guard up, all in the same run — and shows what the combined night is worth. Every caption carries its own point value; the total and the full-field ripple update as you go.
No captions moved yet — nudge a caption above to stack a change.
#
Corps
Gold
Medal
Top 6
The same ten-thousand-season model, but with every move you stacked applied together. Because all your changes land on one corps, the combined swing is exactly the sum of the caption points — no estimate, just more numbers moving at once.
How It Works
The honest version, because the interesting part is the machinery.
The caption math. A DCI total is built from eight captions: the two General Effect scores count full weight, and the six others — brass, percussion, music analysis, color guard, and the two visual captions — count half. When you move a caption to a new rank, we take the score currently sitting at that rank and hand it to your corps, then rebuild their total. Flip to raw mode to dial in an exact number instead.
The re-simulation. That change shifts your corps's projected finals score, and then the same Monte Carlo behind the Road to Indianapolis re-runs the rest of the season — every corps's fitted trajectory, its demonstrated noise, the late-season compression that squeezes the field together. Ten thousand seasons, ranked, counted.
Why the deltas are clean. We run the before and after on the same ten thousand random draws and change exactly one number between them. So when Crown's medal odds jump and Boston's slip, that's a real consequence of the swap, not two noisy simulations disagreeing. The instant estimate you see while dragging is a lighter version of the same math; the confirmed number that lands a beat later is the full ten-thousand-run result.
What it can't see. Whether a book can actually be fixed in three weeks, what it costs somewhere else on the sheet, or how a design ages under an Indianapolis panel. This is a scoreboard thought experiment, not a promise. Treat it the way you'd treat any good bar argument: seriously, and not too seriously.