more succinct benchmark summaries (#6959)

<!--
Thank you for submitting a pull request! We appreciate your time and
effort.

Please make sure to provide enough information so that we can review
your pull
request. The Summary and Testing sections below contain guidance on what
to
include.
-->

## Summary

<!--
If this PR is related to a tracked effort, please link to the relevant
issue
here (e.g., `Closes: #123`). Otherwise, feel free to ignore / delete
this.

In this section, please:

1. Explain the rationale for this change.
2. Summarize the changes included in this PR.

A general rule of thumb is that larger PRs should have larger summaries.
If
there are a lot of changes, please help us review the code by explaining
what
was changed and why.

If there is an issue or discussion attached, there is no need to
duplicate all
the details, but clarity is always preferred over brevity.
-->

Closes: #000

<!--
## API Changes

Uncomment this section if there are any user-facing changes.

Consider whether the change affects users in one of the following ways:

1. Breaks public APIs in some way.
2. Changes the underlying behavior of one of the engine integrations.
3. Should some documentation be updated to reflect this change?

If a public API is changed in a breaking manner, make sure to add the
appropriate label. You can run `./scripts/public-api.sh` locally to see
if there
are any public API changes (and this also runs in our CI).
-->

## Testing

<!--
Please describe how this change was tested. Here are some common
categories for
testing in Vortex:

1. Verifying existing behavior is maintained.
2. Verifying new behavior and functionality works correctly.
3. Serialization compatibility (backwards and forwards) should be
maintained or
   explicitly broken.
-->

---------

Signed-off-by: Will Manning <will@willmanning.io>

Author: lwwmanning

scripts/compare-benchmark-jsons.py

@@ -143,19 +143,6 @@ def ratio_stats(
     }
 
 
-def robust_scale(values: pd.Series | np.ndarray) -> float:
-    """Estimate spread with MAD so outliers do not dominate the noise estimate."""
-
-    array = np.asarray(values, dtype=float)
-    array = array[np.isfinite(array)]
-    if array.size == 0:
-        return float("nan")
-
-    median = np.median(array)
-    mad = np.median(np.abs(array - median))
-    return float(1.4826 * mad)
-
-
 def median_polish(table: pd.DataFrame, max_iterations: int = 10, tolerance: float = 1e-8) -> MedianPolishResult | None:
     """Estimate row and column effects for the log-ratio matrix."""
 
@@ -309,20 +296,16 @@ def build_statistical_analysis(df: pd.DataFrame, threshold_pct: int) -> dict[str
         axis=1,
     )
 
-    # Median polish gives a robust overall shift plus residual-noise estimate.
+    # Median polish gives a robust overall shift estimate.
     log_ratio_table = detail_df.pivot(index="query", columns="combo", values="log_ratio")
     polish = median_polish(log_ratio_table)
-    residual_noise_log_scale = robust_scale(polish.residuals.to_numpy().ravel()) if polish is not None else float("nan")
 
     return {
         "detail_df": detail_df,
         "query_stats": query_stats,
         "systemic_shift_ratio": float(np.exp(systemic_shift_log_ratio)),
         "systemic_shift_std": systemic_shift_std,
         "median_polish": polish,
-        "residual_noise_ratio": float(np.exp(residual_noise_log_scale))
-        if np.isfinite(residual_noise_log_scale)
-        else float("nan"),
     }
 
 
@@ -527,72 +510,46 @@ def main() -> None:
     vortex_df = df3[df3["name"].str.contains("vortex", case=False, na=False)]
     parquet_df = df3[df3["name"].str.contains("parquet", case=False, na=False)]
 
-    geo_mean_ratio = calculate_geo_mean(df3)
     vortex_geo_mean_ratio = calculate_geo_mean(vortex_df)
     parquet_geo_mean_ratio = calculate_geo_mean(parquet_df)
-    overall_performance = (
-        "no data"
-        if pd.isna(geo_mean_ratio)
-        else format_performance(geo_mean_ratio, improvement_threshold, regression_threshold, "overall")
-    )
 
-    summary_lines = [
-        "## Summary",
-        "",
-        f"- **Overall**: {overall_performance}",
-    ]
+    statistical_analysis = build_statistical_analysis(df3, threshold_pct)
+    verdict = build_verdict(statistical_analysis) if statistical_analysis is not None else None
+
+    summary_fields: list[str] = []
+
+    if verdict is not None:
+        summary_fields.append(f"**Verdict**: {verdict['status']} ({verdict['confidence']} confidence)")
+        summary_fields.append(f"**Attributed Vortex impact**: {verdict['impact']}")
+
     if len(vortex_df) > 0:
         vortex_performance = format_performance(
             vortex_geo_mean_ratio,
             improvement_threshold,
             regression_threshold,
             "vortex",
         )
-        summary_lines.append(f"- **Vortex**: {vortex_performance}")
+        summary_fields.append(f"**Vortex (geomean)**: {vortex_performance}")
     if len(parquet_df) > 0:
         parquet_performance = format_performance(
             parquet_geo_mean_ratio,
             improvement_threshold,
             regression_threshold,
             "parquet",
         )
-        summary_lines.append(f"- **Parquet**: {parquet_performance}")
+        summary_fields.append(f"**Parquet (geomean)**: {parquet_performance}")
 
-    statistical_analysis = build_statistical_analysis(df3, threshold_pct)
-    verdict = build_verdict(statistical_analysis) if statistical_analysis is not None else None
     if verdict is not None:
-        summary_lines.extend(
-            [
-                "",
-                "## Verdict",
-                "",
-                f"**{verdict['status']}**",
-                f"- **Attributed Vortex impact**: {verdict['impact']}",
-                f"- **Confidence**: {verdict['confidence']}",
-                f"- **Environment shift**: {verdict['environment_shift']}",
-            ]
-        )
-
-    if statistical_analysis is not None:
-        systemic_shift = format_ratio_change(statistical_analysis["systemic_shift_ratio"])
-        control_sigma = format_ratio_change(float(np.exp(statistical_analysis["systemic_shift_std"])))
-        residual_noise = format_ratio_change(statistical_analysis["residual_noise_ratio"])
-        summary_lines.extend(
-            [
-                "",
-                "## Statistical Summary",
-                "",
-                f"- **Systemic shift ({CONTROL_FORMAT} controls)**: {systemic_shift}",
-                f"- **Control sigma**: {control_sigma}",
-                f"- **Residual noise**: {residual_noise}",
-            ]
-        )
-
-        polish = statistical_analysis["median_polish"]
-        if polish is not None:
-            summary_lines.append(f"- **Median polish overall**: {format_ratio_change(float(np.exp(polish.overall)))}")
-
-    print("\n".join(summary_lines))
+        shifts = f"Parquet (control) {verdict['environment_shift']}"
+        if statistical_analysis is not None:
+            polish = statistical_analysis["median_polish"]
+            if polish is not None:
+                shifts += f" · Median polish {format_ratio_change(float(np.exp(polish.overall)))}"
+        summary_fields.append(f"**Shifts**: {shifts}")
+
+    print("<br>".join(summary_fields))
+    print("")
+    print("---")
     print("")
 
     if statistical_analysis is not None: