I recently - and by recently I mean about an hour ago - figured out the necessary skills to mash together teams and the riders they picked in composite. I posted what I came up with 2013 in some of the other threads, but 2013 is only so interesting. It's only begun. Nothing to learn!
But I realized I could apply the same techniques to 2012 but come up with more interesting results. So I decided to identify what the perfect VDS team looked like in 2012.
It's not actually the perfect VDS team in 2012 - that's an easy exercise using a spreadsheet solver. And I am trying to stay away individual riders - I want to know about the type of rider. I settled on the definition of a perfect team as the following:
- How many of each cost rider (one pointers, two pointers, etc) should I pick?
- How much should I expect each rider in each cost bracket to score (hurdle rate of sorts)?
I used the top 100 VDS teams from 2012 to define the perfect team. So by "perfect," it's more like "Creating a VDS Team with a good likelihood of scoring in the top 100 in 2012"
So, to start with, how many of each cost rider should I pick?
Since my model is not very complex, I have to resolve to using fractions of riders. The rules of the game don't let you pick a fraction of a rider, but use your imagination.
I took the top 100 VDS teams from 2012 and calculated how many riders in each cost bracket they picked. I collapsed the brackets down to ranges (6 to 8, 10 to 12, etc). My results are:
So to create a top 100 team, you should have picked 5.5 one pointers, 0.6 twenty-four+ pointers, etc. Easy to grasp.
But how much should you have expected to earn from riders in each bracket?
What's the hurdle rate you need? I went with efficiency since most people think along those terms. What is the efficiency you should expect per price bracket?
These brackets aren't perfect because of small sample sizes. But I collapsed them to ranges to smooth the data out a bit.
So does my math work out? If you had used the number of riders and achieved my suggested hurdle efficiency rates, my team would have scored 16,775 (calculated before I collapsed the rider costs down into ranges). Which puts my fake team at 39th last year. Just about right.
How much does each type of rider contribute to your final score?
How important are your picks in each bracket?
So if I had posted this a week ago, should you have followed the above guidance in picking your 2013 team? Absolutely not! Every year is a brand new game.