Skip to contentThe 2026 report on adversarial matter intelligence · read now
Free tool

Which lever actually moves your case?

Six dials describe the case — pleading, evidence, forum, precedent. The model shows what they imply: the chance a motion to dismiss ends it, the outcome distribution, and a ranked answer to the question that matters: where does work buy the most?

All free tools
68% survival
Chance the complaint survives a Rule 12(b)(6) motion to dismiss — under your dials, in this transparent model. A neutral complaint sits near 68%.
Where developing the case buys the most (expected-value gain from one notch of improvement):
1. Your evidence
+$17,398 expected value per notch
2. Precedent strength
+$10,150 expected value per notch
3. Pleading strength
+$10,145 expected value per notch
4. Claim severity
+$9,601 expected value per notch
5. Forum favorability
+$8,580 expected value per notch
6. Their evidence
$-18,288 expected value per notch
How this is calculated

This is the Juricratic war room's Bayesian case model, vendored verbatim and pinned by parity tests. It is a transparent generative model — explicit priors and likelihoods you can read, not a trained black box. The outcome distribution is a naive-Bayes log-odds update over four resolutions; every dial's contribution to it is inspectable.

The motion-to-dismiss readout is a logistic sub-model anchored so a neutral complaint survives about 68% of the time (most 12(b)(6) motions do not end the whole case), driven chiefly by pleading sufficiency, then precedent, evidence, and the claim's substance.

The ranked list at the bottom is a sensitivity analysis: the model bumps each quality dial one notch and reports the expected-value gain — the gradient of case value over the case model. That ranking, not any single probability, is the practical output: it says where preparation pays.

Questions
Will a motion to dismiss end my case?
Usually not by itself — most Rule 12(b)(6) motions are denied or granted only in part, which is why the model anchors a neutral complaint near 68% survival. But the probability moves sharply with pleading quality, and surviving with a weakened complaint changes settlement leverage too.
What do the four outcome bars mean?
They are the model's distribution over how the case resolves — plaintiff verdict, settlement, defense verdict, or dismissal — given your dials. They are what-if outputs of a transparent model, not calibrated predictions about your actual case.
What does 'expected value per notch' mean?
The model improves one dial by 0.1 (one notch), recomputes the expected case value, and reports the difference. A big number means that dimension of the case is the binding constraint — that's where marginal effort pays most.
Why does improving 'their evidence' show a negative number?
The list ranks what happens if each dial INCREASES. The defendant's evidence getting stronger hurts the plaintiff's expected value, so its delta is negative — which is itself informative: it quantifies your exposure to their discovery.
Is this legal advice or a prediction?
Neither. Every input is your assumption and every output is a simulation artifact of a readable model. It creates no attorney-client relationship. Use it to structure thinking, not to decide motions.

This tool is a simulation artifact, not legal advice, and creates no attorney–client relationship. Every probability-like input is your assumption — a dial, not a prediction. Your inputs never leave your browser. Verify all deadlines and figures against the rules and orders of your court.

This is one dial. Your case has fifty.

Juricratic models your whole matter as a solvable game — every motion, every deadline, every settlement posture — and shows how the picture moves when the assumptions do.

Request access
simulation, not prediction — not legal advice