updating metric definition

src/components/leaderboard.py

@@ -186,16 +186,23 @@ def render_leaderboard_table(display_df, metric_columns, primary_metric):
             formula_html = """
             <div style="margin: 15px 0;">
                 <p style="margin-bottom: 10px; font-weight: 500;">Formula:</p>
-                <div style="background-color: #111; padding: 15px; border-radius: 5px; text-align: center; margin-bottom: 15px;">
-                    Relative Improvement to Human = max<sub>all runs</sub>((P<sub>agent</sub> - P<sub>baseline</sub>) / (P<sub>human</sub> - P<sub>baseline</sub>)) × 100%
+                <div style="background-color: #111; padding: 20px; border-radius: 5px; text-align: center; margin-bottom: 15px; font-size: 18px; line-height: 1.5; border: 1px solid #333;">
+                    <div style="display: flex; align-items: center; justify-content: center;">
+                        <div style="margin-right: 10px;">Relative Improvement to Human =</div>
+                        <div style="display: inline-block; text-align: center; padding: 0 10px;">
+                            <div style="border-bottom: 1px solid #aaa; padding-bottom: 5px;">s<sub>agent</sub> - s<sub>baseline</sub></div>
+                            <div style="padding-top: 5px;">s<sub>top_human</sub> - s<sub>baseline</sub></div>
+                        </div>
+                        <div style="margin-left: 10px;">× 100%</div>
+                    </div>
                 </div>
                 <p style="margin-top: 10px; font-weight: 500;">Where:</p>
                 <ul style="list-style-type: disc; padding-left: 25px; margin-top: 8px;">
-                    <li style="margin-bottom: 5px;">P<sub>agent</sub> is the agent's test performance</li>
-                    <li style="margin-bottom: 5px;">P<sub>baseline</sub> is the baseline test performance</li>
-                    <li style="margin-bottom: 5px;">P<sub>human</sub> is the human performance benchmark</li>
-                    <li style="margin-bottom: 5px;">The maximum is taken across all experimental runs for a given task-model pair</li>
+                    <li style="margin-bottom: 5px;">s<sub>agent</sub> is the agent's test performance</li>
+                    <li style="margin-bottom: 5px;">s<sub>baseline</sub> is the baseline test performance</li>
+                    <li style="margin-bottom: 5px;">s<sub>top_human</sub> is the top human performance in competition</li>
                 </ul>
+                <p style="margin-top: 10px;">This metric normalizes scores by setting the baseline solution to 0 and the top human solution to 100.</p>
             </div>
             """
             
@@ -206,15 +213,22 @@ def render_leaderboard_table(display_df, metric_columns, primary_metric):
             formula_html = """
             <div style="margin: 15px 0;">
                 <p style="margin-bottom: 10px; font-weight: 500;">Formula:</p>
-                <div style="background-color: #111; padding: 15px; border-radius: 5px; text-align: center; margin-bottom: 15px;">
-                    Absolute Improvement to Baseline = max<sub>all runs</sub>((P<sub>agent</sub> - P<sub>baseline</sub>) / P<sub>baseline</sub>) × 100%
+                <div style="background-color: #111; padding: 20px; border-radius: 5px; text-align: center; margin-bottom: 15px; font-size: 18px; line-height: 1.5; border: 1px solid #333;">
+                    <div style="display: flex; align-items: center; justify-content: center;">
+                        <div style="margin-right: 10px;">Absolute Improvement to Baseline =</div>
+                        <div style="display: inline-block; text-align: center; padding: 0 10px;">
+                            <div style="border-bottom: 1px solid #aaa; padding-bottom: 5px;">s<sub>agent</sub> - s<sub>baseline</sub></div>
+                            <div style="padding-top: 5px;">s<sub>baseline</sub></div>
+                        </div>
+                        <div style="margin-left: 10px;">× 100%</div>
+                    </div>
                 </div>
                 <p style="margin-top: 10px; font-weight: 500;">Where:</p>
                 <ul style="list-style-type: disc; padding-left: 25px; margin-top: 8px;">
-                    <li style="margin-bottom: 5px;">P<sub>agent</sub> is the agent's test performance</li>
-                    <li style="margin-bottom: 5px;">P<sub>baseline</sub> is the baseline test performance</li>
-                    <li style="margin-bottom: 5px;">The maximum is taken across all experimental runs for a given task-model pair</li>
+                    <li style="margin-bottom: 5px;">s<sub>agent</sub> is the agent's test performance</li>
+                    <li style="margin-bottom: 5px;">s<sub>baseline</sub> is the baseline test performance</li>
                 </ul>
+                <p style="margin-top: 10px;">This metric measures the percentage improvement of an agent's performance over the baseline solution.</p>
             </div>
             """