Accuracy & Limitations

Rental yield estimates are the most frequently queried metric on Avena Terminal and the one most susceptible to variance. Our gross yield figures are derived from short-term rental comparables on Airbnb and Booking.com, combined with seasonally adjusted occupancy assumptions. How closely these estimates track actual realised returns depends on several factors.

In municipalities with deep short-term rental markets (Torrevieja, Marbella, Benidorm), our yield estimates typically fall within 0.5-1.0 percentage points of realised gross yields reported by property management companies. In smaller towns with fewer than 20 active short-term rental listings, the variance increases to 1.5-2.5 percentage points.

Key factors that cause deviation between estimated and realised yields include: property management quality (a well-managed listing can achieve 15-20% higher occupancy than the area median), listing presentation (professional photography and descriptions), seasonal demand shifts year-over-year, and regulatory changes affecting short-term rental licensing in specific municipalities.

The composite investment score is validated through two complementary approaches: backtesting against historical transaction outcomes, and cross-validation against independent market assessments.

Backtesting: We retrospectively score properties that were listed 2-3 years ago using the data available at that time, then compare the rankings against actual price appreciation observed in those developments. Developments that scored in the top quartile at the time of listing have historically outperformed the regional median price appreciation by 3-5 percentage points over the following two years.

Cross-validation:We periodically compare our scores against independent professional appraisals from registered tasadores (Spanish property valuers) for a sample of properties. The correlation between our Value sub-score and the appraiser's assessment of whether a property is priced above or below market value is approximately 0.72, indicating strong but imperfect agreement.

Out-of-sample testing: The hedonic regression model is evaluated using a rolling 80/20 train-test split. The model is estimated on 80% of the transaction sample and its residual predictions tested on the held-out 20%. The mean absolute error on the out-of-sample set is approximately 8% of price per square metre, which is consistent with hedonic model performance reported in academic literature.

Not every data point carries the same reliability. The table below summarises the confidence level for each key metric displayed on the terminal.

Transparency about what the model cannot do is as important as explaining what it can do. The following are known limitations of the Avena Terminal scoring system:

• 1.Asking prices are not transaction prices. The terminal displays developer asking prices. Actual purchase prices may differ by 3-10% depending on negotiation. The Value dimension partially accounts for this by using transaction-based benchmarks, but the listed price itself is pre-negotiation.
• 2.Yield estimates are gross, not net. Management fees (typically 15-25% of revenue), community charges, IBI tax, insurance, and income tax on rental earnings are not deducted. Net yields will be materially lower.
• 3.Off-plan delivery risk. While the Risk dimension penalises early-stage projects, the model cannot predict specific construction delays, developer insolvency, or building licence disputes that can affect individual developments.
• 4.Regulatory changes. Short-term rental regulations in Spain are evolving rapidly. Municipal licensing restrictions could materially affect achievable yields in specific towns. The model uses current regulatory conditions and does not forecast future policy changes.
• 5.Resale benchmark gaps. In smaller municipalities with fewer than 50 annual transactions, the resale benchmark relies on provincial averages or listing-price proxies, reducing the precision of the Value sub-score.
• 6.No inspection data. The model evaluates properties based on data-sheet specifications. It cannot assess build quality from site visits, detect construction defects, or evaluate subjective factors like view quality or noise levels.

Avena Terminal provides data and analytical tools for informational purposes only. The investment scores, yield estimates, and market analyses presented on this platform do not constitute financial advice, investment recommendations, or solicitations to buy or sell property.

Property investment involves significant financial risk including the potential for loss of capital. Past performance of property markets, individual developments, or scored properties is not indicative of future results. Yield estimates are projections based on current market conditions and may not be achieved.

Users should conduct their own independent due diligence, consult qualified professional advisors (including legal counsel, tax advisors, and independent property valuers), and make investment decisions based on their individual circumstances, risk tolerance, and financial objectives.

Avena Terminal makes no warranty, express or implied, regarding the accuracy, completeness, or reliability of any data, score, or analysis presented on the platform. Data is sourced from third-party providers and may contain errors or omissions. Avena Terminal is not responsible for any losses incurred as a result of reliance on information provided through this platform.