Testing the Toy

Wildcats5421

Well-known member
I
I understand that point, and you very well may be right. Maybe they are ranked too high. That said, I went and watched them play Friday and, in my humble opinion, they're ranked exactly where they should be.



The upper part of DI is all wonky because those teams are all beating each other. I'm a little surprised by Happy being where they are but, since their loss to Westbrook, they've dominated and they got a huge bump after last week for beating Knox City (ranked #8 at the time) 50-0.

Remember, The Toy doesn't think like we do. It just takes numbers in and spits numbers out. It doesn't see "Westbrook beat Happy so they must be ranked higher than Happy", especially when that happened so many weeks ago. The Toy is definitely not perfect, and I'll work every off-season to see if there were things I could have tweaked to make it better, but picking 42-6 in the playoffs, to this point, is pretty darn good, by any standard, but especially for a computer.
completely understand, just starting a friendly conversation
 

Mike

Administrator
It should but who knows? @Mike can probably shed some light on it but no one can deny Westbrook has played the toughest schedule in any either division this year and that hasn't gotten them any brownie points with the Toy. Still baffles me how week one of the playoffs they 45 GC and get no bump from it, have beaten both teams ranked above them this week and still fall into the #3 spot. Believe me we don't care where we are ranked but trying to figure out how a team who played the toughest schedule this year and decimated GC and BV in the first 2 rounds still sits behind 2 teams they beat is mind boggling to say the least.
Also have to wonder how Abbott who played no one in the top 25 all season stayed in the top 5 of the rankings until the playoffs started and then mysteriously drop to #6?

So, The Toy doesn't put a grade on strength of schedule. It simply doesn't work that way. While beating, or playing close games with, the higher rated teams will get you more points, it's just not something that The Toy takes into consideration, specifically. Here is how it works, put simply.

I feed a list of teams, and ratings to The Toy. It looks like this:

Team A:W.ww
Team B:X.xx
Team C:Y.yy
Team D:Z.zz

Then I feed it a list of scores from each week:

W1:Team A:40:Team B:30
W1:Team C:65:Team D:42

W2:Team A:87:Team C:40
W2:Team B:45:TeamD:0

Then it processes all games, running calculations based on the team's ratings and final scores of the games. It then returns to me a list of all the teams and their new ratings:

Team A:E.ee
Team B:F.ff
Team C:G.gg
Team D:H.hh

Then the teams are ordered by rating and have a rank applied based on their order in this list.

So, as you see, The Toy doesn't know what teams are ranked where, and it doesn't take into account who all each team has played. It only looks at, based on the ratings, what teams should have won and by how much, and calculates new ratings based on that and what actually happened.

Granted, it's much more complicated then I make it sound here and it's much deeper than just calculating one week at a time, or even one team at a time, but you can get the idea.
 

FenceLineGuy

Well-known member
So, The Toy doesn't put a grade on strength of schedule. It simply doesn't work that way. While beating, or playing close games with, the higher rated teams will get you more points, it's just not something that The Toy takes into consideration, specifically. Here is how it works, put simply.

I feed a list of teams, and ratings to The Toy. It looks like this:

Team A:W.ww
Team B:X.xx
Team C:Y.yy
Team D:Z.zz

Then I feed it a list of scores from each week:

W1:Team A:40:Team B:30
W1:Team C:65:Team D:42

W2:Team A:87:Team C:40
W2:Team B:45:TeamD:0

Then it processes all games, running calculations based on the team's ratings and final scores of the games. It then returns to me a list of all the teams and their new ratings:

Team A:E.ee
Team B:F.ff
Team C:G.gg
Team D:H.hh

Then the teams are ordered by rating and have a rank applied based on their order in this list.

So, as you see, The Toy doesn't know what teams are ranked where, and it doesn't take into account who all each team has played. It only looks at, based on the ratings, what teams should have won and by how much, and calculates new ratings based on that and what actually happened.

Granted, it's much more complicated then I make it sound here and it's much deeper than just calculating one week at a time, or even one team at a time, but you can get the idea.
👍🏼👍🏼
 
Top