Code Evolution Chart is a part of the candidate report that helps you understand the candidate's thought process and how they arrived to the submitted solution(s). It is also a very helpful tool to gauge if the candidate potentially cheated during their test.
The Code Evolution Chart aggregates all proctoring data collected during the assessment and combines it with insights into the candidate's coding activity, including the speed and frequency of changes. This generates clear visual indicators that help identify any potentially suspicious behavior.
With these charts, you no longer need to manually review raw proctoring data or individual code snapshots to understand the candidate's activity. Instead, a quick glance at the Code Evolution Chart highlights key moments of potentially suspicious behavior, allowing you to focus only on the most relevant instances.
Helpful examples
Example 1. The candidate was slowly working over the course of the session, without sudden changes. It's unlikely that they tried to cheat.
Example 2. The candidate did leave the IDE, but didn’t make any substantial changes to their code after that.
Example 3. The candidate was leaving the Codility tab, and then coming back with some changes to the code pretty frequently. With both charts going up & down, it looks like the candidate was trying out different approaches, that could’ve been found online.
Example 4. The candidate left Codility for a couple of minutes, and then wrote the entire solution.
It’s highly probable that they found their solution somewhere and typed it manually to ensure that they aren’t caught by the copy-paste detection.
Example 5. The candidate spent considerable amount of time outside of Codility, and then pasted their solution. They may have been working in their own IDE or tried to cheat.
Example 6.The candidate spent a lot of time outside of Codility, tried copying the description, and also, after coming back to the IDE, pasted their solution. It's highly likely that they tried to cheat.
When making decisions, remember to take into account all of the available signals and anti-cheating measures (similarity check, network IP detection, photo ID verification, proctoring snapshots). If you have doubts, we recommend conducting an interview session with the candidate to explain their solutions and how they achieved them.