I know what I mean but maybe I am not explaining myself properly.
You know how you go into the Z stations and they give you those orange chips to put into the charity of choice to donate to? Imagine if that were Phoenix time slots. Then people would tell you by their vote what they felt what was the best time slot and the results would bear out cause of equal choice.
At the moment, you do not have a choice. You largely have to attend the games when the FFA schedule them regardless if that is the best time slot for you. If Phoenix games were ALWAYS scheduled at 5pm Sunday and averaged 20k in crowds, that would not necessarily mean that was the best time to have Phoenix home games in Wellington because other time slots could do better.
In the case of the graph, you have the Sunday 5pm is the standout because thats when the FFA schedule them. The only other one that is comparable to measure against is the Friday 7:30pm slot because it too has a large sample size (16 vs 10) and the crowd sizes are roughly the same. If each time slot had 16 games against them, then you could conclude what was the most popular time slot cause its an even spread and it would smooth out the highs and the lows. If that 1 game on the Wed 8pm had 20k, you could not conclude that it was the most popular time slot. You need to have even distribution of data to draw conclusions and like Tegal said, the one offs and low samples on others means you cannot.