Cash ROI
The normalized all-cash return after the model's property-tax and insurance estimates.
Gem Score methodology
Gem Score converts four normalized financial signals into one consistent weekly comparison. The formula is deterministic, versioned, and intentionally narrower than full underwriting.
Current formula / gem_mvp_v1
Gem Score = 40% Cash ROI + 30% Cap Rate + 20% Gross Yield + 10% Voucher Rent Strength
Every term in the formula is first min-max normalized from 0 to 100 across the entire current Cash Flow Gems score population. If every row has the same value for a signal, that normalized signal receives 50 for the run.
Weight by signal
The normalized all-cash return after the model's property-tax and insurance estimates.
The normalized modeled net operating income divided by purchase price.
The normalized annual HUD voucher-rent benchmark divided by purchase price.
The normalized monthly voucher-rent benchmark within the current scored population.
Implemented financial model
Not included
Deterministic ranking order
After Gem Score, the engine sorts by raw cash ROI, cap rate, gross yield, and monthly voucher rent from higher to lower, then by purchase price from lower to higher. A stable internal record identifier provides the final deterministic tie-breaker.
Worked score-only example
This is a mathematical example, not a property or current market result.
Total Gem Score: 32.0 + 22.5 + 12.0 + 5.0 = 71.5
Interpretation
The score describes a property relative to the current scored population, not a universal threshold of investment quality.
A new listing population or refreshed input can move normalized values even if one property's raw inputs stay unchanged.
Use the score to prioritize review, then replace modeled assumptions with verified property and operating facts.
Questions investors ask
No. Gem Score is a screening and ranking tool. It does not verify property condition, title, taxes after sale, insurance, financing, repairs, vacancy, management, utilities, local compliance, payment standards, rent reasonableness, or inspection outcomes.
Not as an absolute performance measure. The inputs are min-max normalized against the current scored population, so a score can change when the comparison population or underlying data changes. Use the dated score version and raw metrics together.
The current MVP model assumes an all-cash purchase with no financing. Under that simplified model, both metrics use the same modeled net operating income divided by purchase price.
Apply the method