Using LWL's Draft Kit

[BZoods]

Could someone explain what functions were all used in relation to standard deviations to normalize the FSI numbers?

I'm looking to analyze my team for statistically weak performers because I need to clear a couple of spots for returning IR players..

[mbr311]

Do you have the information on an excel spreadsheet like LWLs draft kit? If so we simply used this formula: =STDEV(a1:a9).

Obviously the a1-a9 will depend based on the scoring categories you are comparing but that's the formula we used.

[BZoods]

I already knew how to use STDEV, you just made it seem as if there was more to it than just throwing in that.

I was trying to find a way to use STDEV to produce a new data set that gives each player a value between 1-100 depending on how well they did compared to the average, the stdev, and the rest of the field.

In my preseason rankings I just did =percentrank(C2,C:C,4)*100 to do this (only compared to the rest of the league at a particular position). But it hit a major fault when I came to SHG/SHA. because the majority of the league had neither of these, any player who did had significantly heightened overall percentrank.
I was hoping that using STDEV or a similar formula might be able to normalize this result for me.

[BZoods]

Also: instead of a1:a9, unless you have something in a10:a9999999999999, you can just use a:a and it will take all data from that column. It makes it easier when copying formulas between cells.

[mbr311]

I'm not sure I understand why you think I made more of it than simply applying a STDEV formula in order to truly assess a players final FSI ranking..not sure where I made it out to be more to it than just throwing that in.

I used it simply to show why some players final FSI ranking was inflated - because of one category in my particular case. I didn't want a guy ranked too high because he had a very high FSI rating in SHPs for instance which was inflating his total FSI ranking.

It sounds like you are a whiz with math so I'm sure you'll figure it out. I just used it as another tool to figure out a "true" ranking of a player across all categories as well as using it to score how I drafted or as another tool during the draft - when i drafted a player I would dump him into a spread sheet to see what areas I was doing well in (my scoring categories) so as the draft moved into the later rounds I could target say a guy with high hits because that's where I was lacking. I think you've got it all figured out though, let us know if you find anything interesting!

[BZoods]

"He helped me work up a total FSI and even more importantly a standard deviation formula to determine guys that were way over valued because of 1 or 2 categories"

somehow I just thought that he helped you work out a more specific formula that showed through when taking FSI totals/averages rather than just using STDEV to highlight abnormalities. I'm hoping to find a way to reward players for SHG/SHA (those two categories make for a big difference in my roto league), while not punishing the vast majority of those without them. I'm working on a few other projects as well right now, and I'll share some of my ideas once I refine them.

[mbr311]

Hmm...I see what you're saying BZoods. The standard deviation formula that we used resulted in another number (say 38.8) that I then used in relation to the total FSI ranking. Do you know what I mean by FSI ranking BTW, I never even thought to ask. It's the specific index formula that left wing lock used to judge each players worth relative to your specific scoring format. This was a large part of the draft kit they sent out to those of us that paid.

The STDEV formula unfortunately did not get into specifics if you will - it did not highlight say SHP/SHG as a specific category - rather it highlighted a player being "over valued" in terms of his FSI index score because of inflated numbers in 1 or 2 categories. Now I could go back into the players FSI scores once the STDEV formula was applied and figure out what category caused him to be "inflated." That was rather easy, but time consuming obviously.

It just gave me the ability to look at say Frans Nielson is a different light - the total FSI ranking had him as a top 5 skater..but that was HEAVILY inflated because of his SHPs last year. he had pretty much the maximum FSI score in that category. You take that out of the equation, and he fell into a more "natural" place in my rankings.

I hope that all made sense. I really am not a math whiz lol which is why I was talking with Mike so freqently about this stuff. I am good with stats though. And this was a way for me to better assess a players true worth other than just taking his FSI score into account, and also gave me the ability as I was drafting or as I am still building a team to plug in a players FSI scores for each category and see where I need help in. For example it helped me assess that I wasn't drafting enough guys with hits, so I was able to grab a guy like Callahan. I wasn't drafting enough guys with high PIMs...so that's another way I found it useful.

I'd be very interested to see what you come up with, when you figure out how to reward players for specific categories. The more information the better!

[BZoods]

I did buy the LWL draft kit, so I do know what you mean by the FSI. I found that omitting SHG/SHA from every player produced fairly accurate rankings across the board, but I still want to find a way to implement them.

[mbr311]

I did buy the LWL draft kit, so I do know what you mean by the FSI. I found that omitting SHG/SHA from every player produced fairly accurate rankings across the board, but I still want to find a way to implement them.

Then why not just look at them independently from the rest of the FSI score?

Or are you trying to assess ACTUAL SHP/SHG this season (which of course the FSI is built off projections and last years stats)?

I mean if you just pull out the FSI score for SHP you should be able to find out which guys did well last year and are projected to do well this year. But as we all know, SHP/SHG is probably by far the most difficult category to address with any sort of certainty. I mean not only is it highly situational, but it just doesn't happen a ton.