Accuracy
Independent quarterly bulk test results against actual Land Registry sale prices. No cherry-picking. No simulated data. These are real predictions vs real outcomes.
How we benchmark
We measure against the actual Land Registry sale price — the amount the buyer paid and the seller accepted. A tighter, taller curve means more valuations land closer to the real price.
* Hometrack benchmarks against surveyor valuations, not actual sale prices — a less demanding test.
Industry thresholds
PE10 — the percentage of valuations within ±10% of the actual sale price — is the standard accuracy metric for AVMs. Academic and industry research defines several recognition thresholds.
All figures benchmarked against Land Registry sale prices. Industry thresholds were originally defined against surveyor valuations — a softer test. Our 69.4% overall PE10 against actual sale prices exceeds the Kirchmeyer minimum. For Tier 1 (high-confidence) valuations, PE10 reaches 81.3%, exceeding the Rossini “reasonable” threshold.
Reading the metrics
PE10 (Percentage within 10%)
The proportion of valuations where the model's estimate fell within 10% of the actual sale price. This is the industry-standard accuracy metric for automated valuation models. A PE10 above 80% is considered strong performance for a national AVM.
MdAPE (Median Absolute Percentage Error)
The median percentage by which the model's estimate differs from the actual sale price. Using the median rather than the mean prevents outliers from distorting the figure. Lower is better.
Test count
The number of transactions used in the bulk test. A larger test set produces more statistically robust accuracy measurements. Our quarterly bulk test runs every residential transaction recorded by the Land Registry through our live valuation engine.
Accuracy by segment
Model performance varies by property type, price band, and region. From the Meridian v6.11 Q4-2025 bulk test of 147,188 Land Registry transactions.
By Property Type
| Type | n | MdAPE | PE10 | Bias |
|---|---|---|---|---|
| Flat | 23,817 | 7.2% | 63% | +2.5% |
| Terraced | 47,321 | 5.9% | 69% | +1.7% |
| Semi-Detached | 43,907 | 5.6% | 72% | +1.2% |
| Detached | 32,143 | 6.0% | 70% | +1.1% |
By Price Band
| Band | n | MdAPE | PE10 | Bias |
|---|---|---|---|---|
| <£150k | 21,652 | 9.3% | 53% | +12.6% |
| £150-300k | 59,783 | 5.6% | 72% | +0.8% |
| £300-500k | 43,719 | 5.3% | 75% | -0.8% |
| £500k-1M | 20,176 | 6.3% | 69% | -2.1% |
| £1M+ | 1,858 | 10.0% | 50% | -6.5% |
By Region
| Region | n | MdAPE | PE10 | Bias |
|---|---|---|---|---|
| East Midlands | 10,768 | 5.4% | 72.9% | +1.4% |
| East of England | 15,391 | 5.5% | 73.0% | +1.2% |
| London | 17,306 | 6.0% | 69.5% | +2.4% |
| North East | 7,276 | 7.4% | 61.3% | +1.8% |
| North West | 20,687 | 6.6% | 65.5% | +2.2% |
| South East | 22,906 | 5.4% | 73.3% | +1.1% |
| South West | 17,074 | 5.9% | 70.6% | +1.6% |
| Wales | 7,325 | 6.5% | 65.8% | +1.3% |
| West Midlands | 12,276 | 6.1% | 69.9% | +1.3% |
| Yorkshire | 16,138 | 6.6% | 66.5% | +1.1% |
Accuracy by confidence tier
When the model is confident, accuracy increases substantially. For Tier 1 (high-confidence) valuations — 69.6% of the test set — PE10 reaches 81.3%.
| Tier | n | % of test | MdAPE | PE10 | Bias |
|---|---|---|---|---|---|
| Tier 1 (High) | 70,280 | 69.6% | 4.5% | 81.3% | +0.4% |
| Tier 2 (Medium) | 18,611 | 18.4% | 8.4% | 57.4% | +2.8% |
| Tier 3 (Low) | 12,047 | 11.9% | 9.4% | 52.5% | +3.0% |
What this means for lenders: For roughly 69.6% of properties being transacted, our model delivers 81% PE10 with 4.5% median error and +0.4% bias. These are the properties with strong comparable evidence in active local markets — exactly the properties lenders are most likely to encounter.
Confidence tiers
Each property is assigned an individual Forecast Standard Deviation (FSD) based on comparable evidence, feature completeness, and model prediction uncertainty. The three confidence tiers are a human-readable summary of the underlying granular FSD.
Tier 1 — High Confidence
Strong comparable evidence, standard property type, active local market. The prediction interval is narrow and the model has high certainty. Suitable for lending and investment decisions with appropriate due diligence.
Tier 2 — Medium Confidence
Reasonable comparable evidence with some gaps. The prediction interval is moderate. Useful as a price indication but consider supplementing with local market knowledge.
Tier 3 — Low Confidence
Limited comparable evidence, unusual property, thin market, or significant data gaps. The prediction interval is wide. Treat as indicative only — a formal valuation is recommended.
FSD confidence bands
Each valuation also reports a Forecast Standard Deviation (FSD) — a measure of the model’s prediction dispersion for that property. FSD maps directly to the haircut bands used by rating agencies such as Fitch for RMBS transactions, giving lenders a standardised confidence signal.
Band A
FSD ≤ 0.05
Highest confidence. 2.5% haircut. Narrow prediction interval with dense comparable evidence.
Band B
FSD ≤ 0.10
High confidence. 5% haircut. Strong evidence with reasonable prediction certainty.
Band C
FSD ≤ 0.20
Medium confidence. 10% haircut. Moderate evidence with wider prediction interval.
Band D
FSD > 0.20
Lower confidence. 20% haircut. Limited evidence — consider supplementing with desktop review.
Methodology
These metrics are from an independent quarterly bulk test of 147,188 Q4-2025 Land Registry transactions, run through our live valuation engine. They are not training metrics. Every quarter we test our model against every residential transaction recorded by the Land Registry and publish the full results — including known limitations.
Where our model is weaker
Our model performs best on properties in the £150k–£1M range (PE10 69–75%). It is less accurate at the extremes: properties below £150k are systematically overvalued by approximately 12.6%, and properties above £1M are systematically undervalued by approximately 6.5%. Flats have a mild overvaluation bias of +2.5%. We are actively working to reduce these biases.
Download reports
Want more detail?
Read our overview for a summary of what we do and how we compare, or dive into the technical summary for methodology, feature categories, and regulatory alignment.