Energy infrastructure hiring signal
A dated, source-linked record of what A2 observed, predicted before the outcome, and scored after a fixed window.
| Publisher | Intermesh LLC dba A2 Recruiting |
|---|---|
| Principal | Vince Dufour |
| Method version | 1.0, published 29 August 2026 |
| Coverage | Company-level events in energy infrastructure and adjacent industrial markets |
| Human judgment | Four event axes and each prediction confidence |
| Automated work | Validation, arithmetic, public rendering, and machine-readable output |
| Public formats | Events JSON · Predictions JSON · Calibration JSON |
| Contact | vince@a2recruiting.com · (979) 814-4800 |
Method
A senior hire can be a costly, hard-to-reverse company decision. A2 treats source-linked senior hiring as one indicator of operating conviction, not as proof of future financial performance. Each event is scored by Vince Dufour under method version 1.0. Automation may stage sources and calculate outputs, but it cannot assign the four scores or a prediction confidence.
| Axis | 0 | 3 |
|---|---|---|
| Pedigree delta | Lateral or step down in institutional quality | Left a top-tier platform seat for an unproven venture |
| Compensation risk | Cash compensation protected or improved | Forfeited unvested equity or took a large cash cut for equity |
| Scarcity | Large qualified population | Single digits could do it |
| Irreversibility | Same city and same industry | Industry switch and family relocation |
The raw conviction score is the sum of the four axes, from 0 to 12. It is multiplied by the published seniority weight: C-suite 3.0, EVP 2.5, SVP 2.2, VP 2.0, Director 1.0, and Manager 0.5. The public event file omits person names, prior employers, prior roles, and internal notes.
Prediction rules
- Each prediction records the source capital event, predicted role, stated confidence, and resolution window before the outcome is known.
- The confidence and resolution window are not revised after publication.
- A hit requires a public posting or hire matching the predicted role within the fixed window.
- A miss is recorded when no qualifying public posting or hire is found by the fixed date. The evidence field cites the final public page reviewed, and the internal resolution note records the search performed.
- Every resolved row remains in the public ledger. A difficult or ambiguous case is not deleted.
Probability scoring
The raw Brier score is the mean squared probability error: BS = mean((p - y)^2), where y is 1 for a hit and 0 for a miss. It ranges from 0 to 1, and lower is better.
The reference forecast is a constant probability equal to the observed hit rate in the resolved sample. Its Brier score is BSref = q(1 - q). Brier skill is BSS = 1 - BS / BSref. A positive Brier skill score beats that reference; a negative score trails it. Brier skill is undefined when the reference score is zero.
Exact-confidence bands and trigger-level observed rates always carry their sample size. They are descriptive. A2 publishes no categorical skill conclusion before at least 20 predictions resolve.
Record integrity
The CSV ledgers live in version control. Under normal repository operation, the commit history makes changes timestamped, inspectable, and tamper-evident during diligence. Git history is not literally immutable, so public claims rely on the visible ledger, retained history, cited evidence, and a process rule that fixed fields are never rewritten after the outcome.
Status
Method version 1.0 was published 2026-08-29. The live record currently contains 0 open predictions and no resolved predictions. No calibration conclusion has been earned yet.
Machine-readable files: events, predictions, calibration, and methodology.
Use and privacy
This record is company-level market observation and analysis. It is not investment advice. It contains no candidate contact information, private candidate stance, or confidential statement made by an individual. Source publishers retain their rights in the linked material.