Week 04 · Validated · Published October 14, 2026

Brooklyn leads on value. Philadelphia leads in scale-adjusted forecast accuracy.

ZIP 11201 vs ZIP 19107 · 81 monthly observations per city · October 2026 forecasts.

Research results · Week 04 · Scientific figures

Three views of the evidence

Consistent axes, direct annotations and separated historical versus predictive quantities. Values are derived from the research findings; charts use exact stated metrics.

Figure 01

Next-month condo/co-op forecast

Fig. 1 | October 2026 model estimates, not realized transaction prices. Values refer to the specified ZIP-level ZHVI series.

Figure 02

Projected month-ahead change

Fig. 2 | Model-implied change relative to the latest observed monthly index. Percentage-point comparisons are not realized returns.

Figure 03

Model validation · one-step-ahead MAE

Fig. 3 | 44 expanding-window backtests per city. Dollar MAE is not scale-normalized; do not infer relative predictability without scaling.

Source: Zillow Home Value Index (ZHVI), condo/co-op ZIP series; independently assembled and modeled by ChrionML®. No Zillow endorsement implied.

Limitations and reproducibility

Model selection and reported errors use the same 44 validation windows, February 2023–September 2026, starting with 37 training months. These are validation results, not an independent post-selection test. The empirical residual range is not a calibrated confidence interval.

Full-precision model results · Reproduction script · Brooklyn source CSV · Philadelphia source CSV

Each market was evaluated with 44 expanding-window one-step-ahead backtests. Both selected Holt damped trend. Relative error is approximated by dividing MAE by September 2026 ZHVI, not MAPE. Cross-city value levels do not establish investment merit.

Source: Zillow ZHVI condo/co-op series manually assembled for ZIP 11201 and ZIP 19107. October values are forecasts, not observations.