My problem with the stats are that they draw conclusions based entirely on who starts a game, rather than looking at what happens when either player is on the pitch.
For example, Taylor could come onto the pitch in the 75th minute with us 3-0 down. He could score 4 goals in an inspiring fight back and that table would register it as proof that Mackie should start instead of Taylor. Or the opposite could happen. Taylor could leave the game in the 33rd minute with Oxford 1-0 up, Bristol Rovers could score 3 goals when Mackie is on the pitch, and the stats would show that we shouldn't be starting Taylor.
As I went way too deep into the numbers looking at this last week, I'll keep it at a very high level here (I would prefer to look forwards right now anyway). We score (approximately) every 70 minutes with Mackie on the pitch, and every 50 minutes without him. That is 0.5 goals difference per 90 minutes, which is a big difference.