diff --git a/GCryptLib/artifacts/optimization-performance-log.md b/GCryptLib/artifacts/optimization-performance-log.md
index c019e69..6d9edf0 100644
--- a/GCryptLib/artifacts/optimization-performance-log.md
+++ b/GCryptLib/artifacts/optimization-performance-log.md
@@ -1,17 +1,17 @@
 Old way (400 rounds, no matrix mult): 38.01s
-  400 was the minimum for good diffusion
+  - 400 was the minimum for good diffusion
 
 New way (10 rounds, with matrix mult): 1.16s
-  10 rounds now give sufficient diffusion
-  still using bitsets and strings
+  - 10 rounds now give sufficient diffusion
+  - still using bitsets and strings
 
 With new block class (instead of bitsets) (10 rounds): 0.35s
-  still partially using bitsets and strings
+  - still partially using bitsets and strings
 
 With new block class (6 rounds, still good diffusion): 0.21s
 
 With new templated block class (full block and half block): 0.14s
-  finally no more bitsets
+  - finally no more bitsets
 
 With new sbox, reduction, and expansion function: 0.03s