Mishneh Torah, Sanctification of the New Month 6-8

Hook

Founders obsess over "perfect" data, but perfection is often the enemy of a sustainable ship. You’re either running on real-time feedback (witnesses) or predictive modeling (algorithms). The tension isn't between being right or wrong; it’s between being agile and being structured.

Text Snapshot

"The court would calculate the time of the conjunction... in an exact manner... The first level of these calculations represent approximations... and their accuracy is not great... The essentials of the calculations that are used when a court... does not exist... are referred to as ibbur." (Mishneh Torah, Sanctification of the New Month 6:1–3)

Analysis

Insight 1: Precision vs. Practicality

Rambam distinguishes between the mean movement (the math) and the true position (the reality). Your KPIs are models, not reality. Don't confuse your dashboard for your product-market fit.

Insight 2: Constraints Create Systems

The calendar doesn't just "happen"; it uses "postponements" (lo adu rosh) to ensure the system remains functional for the community. Sometimes you must override "technically correct" data to maintain operational viability (e.g., delaying a release to avoid a weekend support nightmare).

Insight 3: The Power of Standardization

The 1080-unit hour was chosen for its divisibility, not its physical reality. In business, choose metrics that empower your team to communicate and execute easily, not just those that track the most granular data.

Policy Move

The "Calibration Review": Move from a binary "is this metric up or down?" to a quarterly review that asks: "Does this KPI still reflect our 'true' celestial position, or has our business model shifted such that this data is now merely an approximation?"

Board-Level Question

"Are we optimizing for the precision of our internal reporting, or for the actual, observable outcome we promised our customers?"

Takeaway

Your data is a tool for navigation, not a deity. Use models for scale, but keep your eyes on the "sighting" of the market to know when to override the algorithm.

---

KPI Proxy: Model Variance (The delta between your predictive forecast and actual realized revenue). If this variance exceeds 15% for three consecutive months, your "algorithm" is broken.