Hello HBers. (Dan, this post might be mostly for you answer).
So I'm sure you're wondering what I mean by "golden stats". It's simply a placeholder name for something Dan Besbris discussed earlier this season in a podcast. I'll explain in the following. (I apologize, I don't remember which podcast it was)
I recall earlier this year, the Besbris talked about how you needed to average something like 0.9 BLKs per game (BPG) from your entire team to win the BLKs category in a ROTO league.
To elaborate, let's say you have Steph Curry who averages 0.1 BPG, and then you have Clint Capela who averages 1.7 BPG, and the rest of your roster somewhere in the middle of those numbers. You would then take all those BPG stats and then average them together, getting your final team BLK average. If it was 0.9 BPG or over, then you would very likely win BLKs in your ROTO league. That 0.9 BPG number would be the aforementioned, "golden stat".
Now, I found this number crunching to be quite interesting and I began to ponder up these questions..
1. We know the golden stat for BPG. What are the golden stats needed for the other 8 categories? How do you arrive to those stats?
To clarify:
I believe the Besbris used past league info to check how many BLKs the top BLKs team had and that's how he got to his conclusion (I'm not sure though?). If this method is used, is it safe to judge other leagues under the same golden stat number?
2. What is considered a bad/excellent average for each position and for each category?
To clarify:
Categories have different values for every position. Let's use Steph Curry's BPG average as the example again. He averages 0.1 BPG. As a PG, is that considered bad, fair, good, excellent relative to the golden stat of 0.9 BPG? I assume that's "bad".. but how do you come to your conclusion on that and how many BPG would a PG need to get to be considered excellent in relation to the golden stat of 0.9 BPG? On the flipside, what is considered a "bad" BPG for a C?
-----------------------------------------------------------------------------
I apologize if this was a long-winded post. But these are the kinda fantasy topics I find most interesting about the Besbris' Hoop Ball podcast that I don't find anywhere else. Kudos to you, Dan!
This is a thread I'd love to save for future use. If folks could post the winning numbers for their roto leagues and how many teams are in said league, we could reasonably come to a number.
Here are some values from a 10-team league that I saved a couple years ago:
FG% - ~48%
FT% - 82%
3pg - 1.452
Ppg - 17.7
Rpg - 6.32
Apg - 3.85
Stlpg - 1.12
Bkpg - .78
TOpg - 1.6
Obviously, these aren't the same for every 10-team league, but it's a starting point. I'd love to see some 11-team an 12-team projections for roto, too.
I’ve done a similar thing over 4 years for a 14 teamer.
FG 49%
FT 82%
PTS 15.5
REB 6.7
ASS 3.7
3PTM 1.4
STL 1.2
BLK 0.8
TO 1.6
Impressively high marks in some of those for a 14-teamer!
16-17 Season 10-team Winning Powerball Numbers:
FG% - 0.484
FT% - 0.833
3PT - 1.51
PTS - 18.38
REB - 6.50
AST - 4.08
STL - 1.14
BLK - 0.93
TO - 1.54
16-17 Season 13-team Winning Mega Millions Numbers:
FG% - 0.497
FT% - 0.839
3PT - 1.98
PTS - 17.95
REB - 7.28
AST - 3.86
STL - 1.20
BLK - 0.79
TO - 1.53
I have the winning lottery numbers for the prior 3 years saved somewhere as well but can't for the life of me find it. I'll throw them in here if I do.
Two notes I'd like to make:
1) A few years ago I remember the winning 3PT number was about ~1.3 as opposed to the ~1.5 it is today, which is indicative of the way the game is shifting to the perimeter.
2) The winning number for each stat will be vastly different across leagues (even of the same size) as you will inevitably find teams that will use punt strategies and concentrate on fewer cats, often artificially inflating them. An example is my 13-teamer above where the leader in 3PTs punted big man stats and had shooters up and down his roster. There's no way you should have to shoot 1.98 3PTs per game to win that cat normally.
If I were to throw out a quick hypothesis (and I haven't done the math to back this up), I'd say that if you took the top X players in the NBA (where X is the size of your league, 12-team x 13-player roster = 156 players) and found the average of each cat over that player sample (bear in mind you'd need to take into account weighted FG% and FT%). Then you could find the standard deviations across each cat as well. And I'd guess that the winning numbers (or there abouts) would be a certain number of standard deviations above the mean, which would be the same across all cats. Just a guess though. Would be interesting to find out.
This thread is SO good. Like, easily my favorite forum thread to date. Even more than the one I made for people to just talk to me.
Wow that is some great info!
These numbers are mind-blowing and actually very unexpected. Regardless of the amount of teams/players rostered, the % cats are much, much higher than I expected. From the info we got here, the minimum for the FG% golden stat is 48%, and I always felt like 45-46% would be the golden stat. Same goes with the FT%. I expected 80% FT% on the dot, but it appears to be 82% minimum. PPG also is much higher than I expected, especially for the 13-14 team leagues.
Despite all this great info, and I am definitely grateful.. The meat of my pondering still remains, which is the 2nd question of my original post. What is considered a bad/excellent average for each position, for each category and how do we get to that conclusion?
I must know! haha
Hey leftee. I think you're definitely low on the FG% assumption. The simple average FG% of the top 200 players in the NBA is already 46.5% so you'd definitely need to be higher than that to win the cat. Also bear in mind that you don't need to win every cat to win your league. You just need to be very good in most cats and that should be sufficient.
With regards to your second question, I'm not sure it's particularly relevant. I personally don't care which position my stats come from as long as my overall team has enough of each cat at the end of the day. For example, you might find yourself without any traditional PGs, but if you have LeBron and Draymond then you're still going to be ok in assists overall. Or consider Jokic, a Center who doesn't block as many shots as other Cs. He is still elite because he derives his value in other ways. And this will continue to be a trend I think as the game moves more towards positionless basketball.
If you really wanted to do the study though then all you have to do is do the same process I described earlier, except instead of using the whole pool of players, you limit the pool via position. Then you can calculate the z-scores (number of standard deviations above/below the mean) for each cat for each player in that position. Hope that helps
PS. I'm pretty sure Ryan Knaus over at Rotoworld does this kind of study each season.
