diff --git a/src/reporting/render_report/report-template.qmd b/src/reporting/render_report/report-template.qmd
index 76614d055..c09c5a21a 100644
--- a/src/reporting/render_report/report-template.qmd
+++ b/src/reporting/render_report/report-template.qmd
@@ -585,15 +585,15 @@ method_details <- purrr::map_dfr(method_info, function(.method) {
 metric_details <- purrr::map_dfr(metric_info, function(.metric) {
   data.frame(
     metric = .metric$name,
-    metric_label = .metric$label
+    metric_label = .metric$label,
+    metric_maximize = .metric$maximize
   )
 }) |>
   dplyr::arrange(metric)
 
-metric_maximize <- purrr::map_lgl(metric_info, "maximize") |>
-  purrr::set_names(metric_details$metric)
-
-metric_reverse <- names(metric_maximize)[metric_maximize == FALSE]
+metric_reverse <- metric_details |>
+  dplyr::filter(metric_maximize == FALSE) |>
+  dplyr::pull(metric)
 
 scores <- purrr::map_dfr(task_results$results, function(.result) {
   if (!.result$succeeded) {
@@ -1181,11 +1181,18 @@ funkyheatmap::funky_heatmap(
 
 ::: {.callout-note}
 This table displays the scaled metric scores.
+After scaling, higher scores are always better and any missing values are set to 0.
 The "Overall" dataset gives the mean score across all of the actual datasets.
 
+Raw scores are also provided to help with understanding any issues but should not be used to compare performance.
+
 Sort and filter the table to check scores you are interested in.
 :::
 
+::: {.panel-tabset}
+
+### Scaled scores {.unnumbered .unlisted}
+
 ```{r}
 #| label: results-table
 #| eval: !expr has_controls
@@ -1246,3 +1253,69 @@ reactable::reactable(
   searchable = TRUE
 )
 ```
+
+### Raw scores {.unnumbered .unlisted}
+
+```{r}
+#| label: results-table-raw
+#| eval: !expr has_controls
+dataset_scores_raw <- complete_scores |>
+  dplyr::select(dataset, method, metric, value) |>
+  tidyr::pivot_wider(
+    names_from = metric,
+    values_from = value
+  )
+
+table_data_raw <- dataset_scores_raw |>
+  dplyr::mutate(
+    dataset = factor(
+      dataset,
+      levels = dataset_details$dataset,
+      labels = dataset_details$dataset_label
+    ),
+    method = factor(
+      method,
+      levels = method_details$method,
+      labels = method_details$method_label
+    )
+  ) |>
+  dplyr::relocate(dataset, .after = method) |>
+  dplyr::arrange(dataset, method)
+
+reactable::reactable(
+  table_data_raw,
+
+  columns = c(
+    list(
+      method = reactable::colDef(
+        name = "Method",
+        sticky = "left"
+      ),
+      dataset = reactable::colDef(
+        name = "Dataset",
+        sticky = "left",
+        style = list(borderRight = "2px solid #999"),
+        headerStyle = list(borderRight = "2px solid #999")
+      )
+    ),
+    purrr::map(metric_details$metric_label,
+      function(.metric_label) {
+        reactable::colDef(
+          name = .metric_label,
+          format = reactable::colFormat(digits = 3)
+        )
+      }
+    ) |>
+      purrr::set_names(metric_details$metric)
+  ),
+
+  highlight = TRUE,
+  striped = TRUE,
+  defaultPageSize = 25,
+  showPageSizeOptions = TRUE,
+  filterable = TRUE,
+  searchable = TRUE
+)
+```
+
+:::