5 premo FWDS

[Marstar]

what the shower, I thought I responded to this.

Basically, as we showed, you can use different numbers to prove your point, whatever side of the argument is. It's better to use real life examples rather than made up stats in some cases.

Like stanton vs Joel selwood.

it all depends on how inconsistent they are.

You are totally correct. As we've already pointed out. Standard deviation and Sample size CAN make the higher average person lose.

It's not hard to find many, many, examples when the odds are about 48:52 as is the case with a 10% ppg difference between 2 players.

But logically, mathematically and statistically it's common sense that higher is higher and lower is lower.

[Samm79]

what the shower, I thought I responded to this.

Basically, as we showed, you can use different numbers to prove your point, whatever side of the argument is. It's better to use real life examples rather than made up stats in some cases.

Like stanton vs Joel selwood.

I'll use dt for this since their numbers are more comparable (stanton averages slightly more). The name I write down wins the round they both played 20 games, albeit, missed certain rounds etc. But I cbf finding what rounds they missed, so i'll just match their numbers up in order :P


selwood/stanton/winner

      
114   153   stanton
104   116   stanton
106   93   selwood
130   175   stanton
69   77   stanton
106   193   stanton
136   140   selwood
100   124   stanton
62   93   stanton
122   105   selwood
103   89   selwood
119   115   selwood
105   111   stanton
97   74   selwood
95   85   selwood
113   84   selwood
125   62   selwood
88   72   selwood
67   61   selwood
101   109   stanton

Stanton: 9
Selwood: 11

Stantons average: 106.6
Selwoods average: 103.1


it all depends on how inconsistent they are.

I cannot agree with this, this argument, it's like flipping a coin 50 times and getting 30 tails and 20 heads, and then making an argument that the coin is favouring tails... You need a bigger sample size. Flip that coin 5,000 times and you will get the 3:2 ratio much less often (almost never).

Compare 10 pairs of players with almost identical averages but differing SD over 3 seasons and then you may have a statistically significant sample size, otherwise it is (statistically) insignificant information.

[Sabretooth Tigers]

 :o
In reply to the original question (back on page 1) I've gone 4 keepers, 1 possible break out, and 3 cheapies. I like the idea of filling the forward line and  leaving the defence til last.  ;)