Go beyond formation energy — measure real-world synthesizability.
Practical Stability Scorer
Formation energy alone misses half the story. The Practical Stability Scorer combines DFT thermodynamics, experimental synthesis records from 38K recipes, cross-database agreement, MC3D validation, and COD crystal structure confirmation into a single deployable score. Endpoint: GET /v1/discovery/practical-stability.
Key outcomes
What goes in
Inputs this feature expects
- Candidate compositions from search, generation, or literature
- Optional weighting for experimental vs. computational evidence
- Target application context to adjust stability thresholds
What comes out
Outputs your team can act on
- Composite practical stability score with per-dimension breakdown
- Synthesis evidence count and source provenance
- Ranked candidate shortlist for downstream simulation or lab work
Workflow
How teams use Practical Stability Scorer
Aggregate the evidence layers
DFT hull distance, synthesis count, cross-source agreement, MC3D, and COD data are pulled from the governed warehouse in one pass.
Compute the composite score
Each dimension is normalized and weighted to produce a deployable stability signal that reflects real-world synthesizability, not just computation.
Prioritize with confidence
High-scoring candidates advance to simulation or synthesis; low-scoring candidates are flagged for further investigation or elimination.
Best fit
Where this feature adds the most leverage
- Teams burning DFT budget on unstable or un-synthesizable compositions
- Discovery programs that need experimental corroboration alongside computational data
- Portfolio reviews where practical feasibility matters as much as predicted performance
Related features