Cleanup

Author: ekoutanov

brumby/src/harville.rs

@@ -358,7 +358,7 @@ mod tests {
     }
 
     #[test]
-    fn harville_summary_3x2_condensed_without_scratchings() {
+    fn harville_summary_2x3_condensed_without_scratchings() {
         const WIN_PROBS: [f64; 3] = [0.6, 0.3, 0.1];
         const RANKS: usize = 2;
         let probs = Matrix::from(

brumby/src/market.rs

@@ -133,7 +133,6 @@ impl Market {
         let overround = prices.invert().sum::<f64>() / fair_sum;
         let est_rtp = 1.0 / overround;
         let initial_k = 1.0 + f64::ln(est_rtp) / f64::ln(prices.len() as f64);
-        // println!("fit_power: initial_k: {initial_k}");
         let outcome = opt::univariate_descent(
             &UnivariateDescentConfig {
                 init_value: initial_k,
@@ -152,7 +151,6 @@ impl Market {
                 (sum - 1.0).powi(2)
             },
         );
-        // println!("fit_power: outcome: {outcome:?}");
 
         let probs = prices
             .iter()
@@ -187,8 +185,6 @@ impl Market {
                 let mut sum = 0.0;
                 for &price in &prices {
                     let uncapped_scaled_price = 1.0 + (price - 1.0) / d;
-                    // let capped_scaled_price =
-                    //     cap(uncapped_scaled_price, MIN_PRICE, MAX_PRICE);
                     sum += 1.0 / uncapped_scaled_price;
                 }
 

brumby/src/opt.rs

@@ -49,25 +49,18 @@ pub fn univariate_descent(
     }
 
     let (mut value, mut step) = (config.init_value, config.init_step);
-    // println!("initial value: {value}, residual: {residual}, step: {step}");
     let (mut optimal_value, mut optimal_residual) = (value, residual);
-    // let mut boost = 1.0;
-    // let mut gradient: f64 = 1.0;
     while steps < config.max_steps {
         steps += 1;
-        let new_value = value + step;/* * boost*/ // * f64::min(gradient.abs(), 100.0);
+        let new_value = value + step;
         let new_residual = loss_f(new_value);
-        // let gradient = (new_residual - residual) / (new_value - value);
-        // println!("iterations: {iterations}, value: {value}, residual: {residual}, step: {step}, new_value: {new_value}, new_residual: {new_residual}");
 
         if new_residual > residual {
             step = -step * 0.5;
             if step.abs() < config.min_step {
                 break;
             }
         } else if new_residual < optimal_residual {
-            // boost = f64::min(gradient.abs(), 10.0);
-            // println!("optimal_residual: {optimal_residual}, new_residual: {new_residual}, boost: {boost}, diff: {}", optimal_residual - new_residual);
             optimal_residual = new_residual;
             optimal_value = new_value;