i have two datasets which i want to find how much they are correlated.
the datasets represent the results of matches of two teams, where 1 represents a win, 0 represents a draw and -1 represents a loss.
e.g. for 5 games
team1 = [1,1,0,-1,0]
team2 = [0,1,0,1,0]
calculating the pearson correlation coefficient is fine till the point where one team won the last 5 games, hence a constant array, e.g.
team1 = [1,1,1,1,1]
In this case the pearson correlation coefficient will be undefined regardless of what team2 did.
I find this weird, because if the team2 also won most of the 5 games, the correlation should be close to 1 actually, not undefined.
and vice versa, if team2 lost most of their matches, the correlation should be close to -1 based on my understanding.
am I doing something wrong here? or my data needs another method to find how strong the relation between the datasets?
Thank in advance
