diff --git a/Feedback - Common mistakes from 2019.md b/Feedback - Common mistakes from 2019.md
index d4b3c99..933aabf 100644
--- a/Feedback - Common mistakes from 2019.md	
+++ b/Feedback - Common mistakes from 2019.md	
@@ -1,24 +1,3 @@
-# Background
-
-- Only giving the "popular science" definition of TSP. You need to be more precise/technical, using appropriate notation.
-- Defining multiple versions of TSP and not being clear on which version you are studying.
-- Not defining the actual variant of TSP that you worked on. For example, **Euclidean TSP** specifies "vertices on the plane" using (x,y) coordinates, and the distance between vertices is then calculated using Pythagoras Theorem. You then also need to check this version is still NP-hard, and cite a reference to confirm this fact.
-- The given definition of TSP suggests optimising **time**, when it should be optimising **cost** (sum of weights).
-- Confusing "problem definition" with "algorithm/solution method".
-   - **Problem**: gives a description of the encoding of a problem, and then asks for property.
-      For TSP: this is "given a weighted complete graph … find the shortest cycle."
-   - **Algorithm**: a step by step mechanism that we can use to program a computer to use the problem's description and recover the required property.
-      This is usually described using pseudocode, and then implemented in real code.
-- It is more accurate to refer to TSP as "computational (optimisation) problem" rather than "mathematical problem".
-- Define number of nodes in graph to be N but then use n later. (Symbols are case sensitive.)
-- In the literature, a graph G=(V,E) is specified by sets V of vertices and E of edges. V and E are **not** numbers - they are sets.
-- Using N for both the "set of nodes" and its size.
-- Mistakes in understanding the meaning of NP-hard.  
-- Mistakenly stating that the **optimisation version** of TSP is in NP. It's not because it is not a **decision problem**. NP only contains **decision** problem - nothing else!
-- Justifying NP-hardness by reducing TSP to another NP-complete problem, when it should be the other way round.
-- Stated that exhaustive search is inefficient but did not say why.
-- Exhaustive search costs O(n!), not O((n-1)!).
-
 # Methodology
 
 - Methodology not stated explicitly in the text, e.g. parameters of generation, what is measured and what is averaged.
@@ -74,12 +53,6 @@
 - Time without units (e.g. seconds or minutes?)
 - Accuracy without units (e.g. percentage?)
 
-# Reflection
-
-- Did not identify any transferable skills or knowledge gained from the project.
-- Identified some gains but did not specify exactly how they will help in the future.
-- Identified some weaknesses but did not give concrete suggestions on how to improve them.
-
 # References
 
 - Not referencing the first paper that proved a fact, e.g. that TSP is NP-hard.