The candidate notes that for a decade, the candidate who supported reform won in a region. This is a correlation. For a couple of instances, it has been the case that the pro-reform candidate won in the northeast. Do we know that their pro-reform stance was the reason why they won the northeast? Of course not. Maybe the northeast simply tends to vote more liberal or conservative than the rest of the region, and the candidate who aligns with that ideology also has happened to support reform.
From this, they conclude that if they support reform, they will win the northeast. This is not true. There could have been alternative causes for the correlation we saw.