|
Post by alantech on Sept 17, 2014 22:47:09 GMT -5
I try to develop and refine metrics of hitting and other volleyball skills on my site VolleyMetrics ( volleymetrics.blogspot.com/), now in its eighth year of operation. If anyone wants to send me a write-up, I'll consider publishing it, crediting you as a guest contributor.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2014 9:58:48 GMT -5
Since dividing by zero equals zero... That is not correct. Dividing anything (including zero) by zero is mathematically undefined; it is not zero. Loosely speaking, people sometimes say it is "infinity", but that is not correct either. What a mathematician would say, for example, is that the limit of 1/x as x approaches zero is unbounded. I will concede to your statement about dividing by zero. For the sake of this statistic when you enter "X/0" into a formula the result given is zero, hence my original statement.
|
|
|
Post by newenglander on Sept 22, 2014 11:02:59 GMT -5
I always like looking into deeper stats but I don't think this stat would be useful for comparing players on 2 different teams, just for players on the same team (and for me it's still a "maybe"). Here's why: - Two hitters on two different teams make the same "0" hit (no kill or error), hitter A has a great blocking and defensive team and hitter B is the only one on their team over 5'7" tall.
This comparison would have hitter A ranking higher because her team and block defend better.
I think hitting efficiency works because it's against a variety of opponents whereas this stat does vary opponents but each hitter has a more static lineup around them that will influence their number in one direction or another.
|
|
|
Post by johnbar on Sept 22, 2014 16:42:48 GMT -5
Sorry for all the negative comments. I certainly think there is room to explore new statistical quantities for rating players. I just don't think this one is well constructed. I have wondered about using (kills)*(hitting percentage), but I have never explored whether that produces meaningful or useful numbers.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2014 18:53:30 GMT -5
Sorry for all the negative comments. I certainly think there is room to explore new statistical quantities for rating players. I just don't think this one is well constructed. I have wondered about using (kills)*(hitting percentage), but I have never explored whether that produces meaningful or useful numbers. I'm certainly open to criticism as long as it is useful. There's nothing worse than that parent who bashes your coaching style and decisions yet offers no solutions (you didn't do that, just saying that annoys me). I will disagree that my argument for the statistic is not well constructed. One may argue that my premises may be wrong, useless, and completely irrelevant, however I believe my defense of the stat is justified based on the data I provided. I punched in some numbers with (k)*(HP) and saw how they can be affected, however I'm not sure the reason for that equation. I added a step and did (k-e)*(HP) and got similar results, however still not sure (yet without much research or thought) of the usefulness of that formula. I would be willing to explore that some more when time permits. I'm also open to hearing any other suggestions concerning my stat or others.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2014 21:09:05 GMT -5
My thoughts to the original poster. Yes, middle hitters generally get better sets than outside hitters in total. We all know that. The hitting percentage is the same for everyone. So all everyone does is adjust the hitting percentage that is considered good. So, an OH hitting .300 is far better than an MH hitting .300. I always adjust for position. A great OH hitting percentage is .300, but for a MH it is closer to .400. Yes?
|
|
|
Post by FOBRA on Sept 22, 2014 21:21:37 GMT -5
As mentioned before, just tacking on HP is so slight that this is basically just a K/E ratio. It doesn't really address the value of volume hitting for OHs, as in your stat the middles still get a number 1 or 2 points above even the most beastly efficient OHs (Robinson and Vansant).
What you really want to do is try a formula that account for changes in attempts. Something like (K-E)/Sqrt(Attempts) might be an interesting place to start. In the PAC-12 I'd imagine low hitting percentage players like Kingdon/Brenner/Gardner would start to get in the ballpark of high efficiency middles, while still propping up a high volume/high efficiency player like Vansant.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 24, 2014 13:26:26 GMT -5
My thoughts to the original poster. Yes, middle hitters generally get better sets than outside hitters in total. We all know that. The hitting percentage is the same for everyone. So all everyone does is adjust the hitting percentage that is considered good. So, an OH hitting .300 is far better than an MH hitting .300. I always adjust for position. A great OH hitting percentage is .300, but for a MH it is closer to .400. Yes? Yes, and I agree with you. This statistic was generated with full thought to the OH. I compared k/e+hp and HP to note the similarities and differences each stat provides, both for M and OH. My intent was to find a number that showcases positive play regardless of position. Naturally, middles should always have a higher efficiency so creating that stat might literally be impossible. The reason I do like this stat for OH is because it represents points won vs points lost, but even at that it has its shortcomings (ex: 11k/11e= 1; statistically, this should be considered neither positive or negative since points scored and points lost are identical).
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 24, 2014 13:30:14 GMT -5
As mentioned before, just tacking on HP is so slight that this is basically just a K/E ratio. It doesn't really address the value of volume hitting for OHs, as in your stat the middles still get a number 1 or 2 points above even the most beastly efficient OHs (Robinson and Vansant). What you really want to do is try a formula that account for changes in attempts. Something like (K-E)/Sqrt(Attempts) might be an interesting place to start. In the PAC-12 I'd imagine low hitting percentage players like Kingdon/Brenner/Gardner would start to get in the ballpark of high efficiency middles, while still propping up a high volume/high efficiency player like Vansant. This is not something that I had even considered and I need to do more research on the sqrt aspects. However, I took a look at some numbers with your suggestion and really liked what I saw. My goal was to develop a position-neutral stat for hitters and the struggle has been finding a way to balance M and OH based on OOS hitting and attempts.
|
|
|
Post by s0uthie on Sept 24, 2014 13:43:14 GMT -5
In my head, the way I would tweak hitting efficiency is to weight it based on the effectiveness of the preceding pass. I would need to do some regressions to find a proper weighting but the traditional 1,2,3 pass would a decent starting point.
|
|