diff --git a/src/main.rs b/src/main.rs
index d2953e7..4678d59 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -19,6 +19,10 @@ struct Cli {
     /// Machine output
     #[arg(short, long, default_value_t = false, global = true)]
     machine: bool,
+
+    /// Verbose output
+    #[arg(short, long, default_value_t = false, global = true)]
+    verbose: bool,
 }
 
 #[derive(Subcommand, Debug)]
@@ -79,7 +83,7 @@ pub fn main() {
                     let b = num_bigint::BigInt::from_str(&mod_exp_args.base.as_str()).expect("could not make bigint");
                     let e = num_bigint::BigInt::from_str(&mod_exp_args.exp.as_str()).expect("could not make bigint");
                     let f = num_bigint::BigInt::from_str(&mod_exp_args.field.as_str()).expect("could not make bigint");
-                    let result = modular_exponentiation(b.clone(), e, f);
+                    let result = modular_exponentiation(b.clone(), e, f, args.verbose);
                     if args.machine {
                         println!("{}", result)
                     }
diff --git a/src/modular_exponentiation.rs b/src/modular_exponentiation.rs
index 19f76cb..588b38c 100644
--- a/src/modular_exponentiation.rs
+++ b/src/modular_exponentiation.rs
@@ -1,9 +1,7 @@
 #![allow(dead_code)]
 
-use num_bigint::BigInt;
+use num_bigint::{BigInt, BigUint};
 use num_traits::ToPrimitive;
-use num_traits::FromPrimitive;
-use pyo3::exceptions::PyException;
 use pyo3::exceptions::PyValueError;
 use pyo3::prelude::*;
 
@@ -16,48 +14,64 @@ pub fn calc_exp_in_field_lib(
 }
 
 /**
- *  modular exponentiation algorithm with big numbers.
- *
- *  Umwandlung des Exponenten k in die zugehörige Binärdarstellung.
- *  Ersetzen jeder 0 durch Q und jeder 1 durch QM.
- *  Nun wird Q als Anweisung zum Quadrieren und M als Anweisung zum Multiplizieren aufgefasst.
- *  Somit bildet die resultierende Zeichenkette von links nach rechts gelesen eine Vorschrift zur Berechnung von x k . 
- *  Man beginne mit 1, quadriere für jedes gelesene Q das bisherige Zwischenergebnis und 
- *  multipliziere es für jedes gelesene M mit x .
+ * square and multiply
  */
 pub fn modular_exponentiation(
     base: BigInt,
-    orig_exp: BigInt, 
-    field: BigInt) -> BigInt {
-    let binary_repr = orig_exp.to_bytes_be();
+    exp: BigInt, 
+    field: BigInt,
+    verbose: bool) -> BigInt {
+    if verbose {
+        println!("args:\nbase {base}\nexp {exp}\nfield {field}\nverbose {verbose}");
+    }
+    let mut instructions: Vec<bool> = bigint_to_bools(exp);
+    // remove the signing bit
+    if verbose {
+        println!("pre instructions {:?}",instructions);
+    }
 
-    let instructions: Vec<bool> = bytes_to_bools(&binary_repr.1);
+    instructions.reverse();
+    if verbose {
+        println!("instructions {:?}",instructions);
+    }
 
-    let mut exp = BigInt::from(1);
+    let mut res = base.clone();
     for instr in instructions {
-        if instr {
+        if verbose {
+            println!("current res: {res}");
+        }
+        if !instr {
             // square
-            exp = (exp.pow(2) * &base) % &field;
+            if verbose {
+                println!("square");
+            }
+            res = res.pow(2) % &field;
         }
         else {
             // square and multiply
-            exp = exp.pow(2) % &field;
+            if verbose {
+                println!("square and multiply");
+            }
+            res = (res.pow(2) * &base) % &field;
         }
     }
 
-    return exp;
+    return res;
 }
 
 #[pyfunction]
 #[pyo3(name="modular_exponentiation")]
+#[pyo3(signature=(base, orig_exp, field, verbose = false))]
 pub fn py_modular_exponentiation(
     base: i128,
     orig_exp: i128, 
-    field: i128) -> PyResult<u128> {
+    field: i128,
+    verbose: bool) -> PyResult<u128> {
     let big_res = modular_exponentiation(
         BigInt::from(base), 
         BigInt::from(orig_exp), 
-        BigInt::from(field)
+        BigInt::from(field),
+        verbose
         );
     let res = big_res.to_u128();
     match res {
@@ -71,17 +85,46 @@ pub fn py_modular_exponentiation(
 
 }
 
-// Vec<u8> to Vec<bool> ( binary representation interpreted otherwise )
-fn bytes_to_bools(bytes: &Vec<u8>) -> Vec<bool> {
+/// Dont use this buggy mess
+pub fn binary_exponentiation(base: BigInt, exp: BigInt, verbose: bool) -> BigInt {
+    if exp.clone() < BigInt::from(0) {
+        return binary_exponentiation(1/&base, -exp, verbose);
+    }
+    else if exp.clone() == BigInt::from(0) {
+        return BigInt::from(1);
+    }
+    else if exp.clone() % 2 == BigInt::from(0) {
+        return binary_exponentiation(&base*&base, &exp/2, verbose);
+    }
+    else if exp.clone() % 2 == BigInt::from(1) {
+        return binary_exponentiation(&base*&base, (&exp-1)/2, verbose);
+    }
+    else {
+        panic!("I don't know how we got here")
+    }
+}
+
+fn bigint_to_bools(item: BigInt) -> Vec<bool> {
     let mut result: Vec<bool> = Vec::new();
-    for byte in bytes {
-        for c in format!("{:b}", byte).chars() {
-            result.push(c == '1');
+    let mut modul : BigInt;
+    let mut smaller = item;
+    loop {
+        if smaller < BigInt::from(2) {
+            break;
+        }
+        modul = &smaller % BigInt::from(2);
+        smaller = &smaller / BigInt::from(2);
+        if modul == BigInt::from(1) {
+            result.push(true);
+        }
+        else {
+            result.push(false);
         }
     }
     result
 }
 
+
 fn dump_bin(bytes: &Vec<u8>) {
     for byte in bytes.iter() {
         println!("{:#08b}\t| {:#02x}", byte, byte);