|
Post by mikegarrison on Jan 27, 2019 13:07:30 GMT -5
Is this what you think you're doing? That's just laughable. You're confirming your own bias because you manipulated the math (15-2-1=12). You're claiming that Stanford's aces were the difference, desperately looking for a real signal, ignoring every example and fact I sent your way. It's interesting that the posters who endorsed the analysis also each stated that the analysis confirmed what they believed or concluded already. What better example of confirmation bias can one ask for? You misunderstood what I was saying. I performed a rather quick and dirty analysis and came to a tentative conclusion. Joe performed a more exhaustive analysis and came to pretty much the same conclusion. I have plenty of experience of the quick analysis missing something important that a later analysis finds, so I was pleased that Joe's analysis found (among other things) the same result. That's not "confirmation bias", that's just "confirmation". There is a difference.
|
|
|
Post by Scipio Aemilianus on Jan 27, 2019 13:10:10 GMT -5
in the third— in their “blowout.” So 25-16 in a National Championship match isn't a "blowout". Interesting use of quotes there. In your honest opinion, what is your reasoning for why Stanford won and Nebraska lost?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 14:04:02 GMT -5
These arguments seem to miss the forest for the trees. Overall this was a pretty close match and the stats are very close. Stanford clearly served better and that was critical, particularly in that first set. Nebraska hit a bit better (not a lot) but shhhhh is quite right that some of that advantage was concentrated in set 4. That overall answer is that this was a close match with small differences. Good LORD - you guys are arguing 'round-and-round 'bout an argument; QUITE flawed at that; from a VB-idiot (m.g.) who: 1) Never played; 2) Never coached; 3) never ref'ed; plus, 4) has spent over 37,000 posts bragging about his engineering expertise-prowess. LMFAO! What bigjohn043 stated above; which corroborates w/ what Joe, shh, and myself have stated here (can NOT believe I have 'liked' as many of sh's posts as I have in this thread - a record!); is that VB IS an i9diosynchratic sport: broken up into 25-point sets. Dr. Strange, Joe, shh and I have ALL been saying the same thing, 'though in different ways. Garrison, and a few others here, be CLUELESS.
|
|
|
Post by Scipio Aemilianus on Jan 27, 2019 14:22:40 GMT -5
is that VB IS an i9diosynchratic sport: broken up into 25-point sets. So this is what you're saying? It's as simple as this? Volleyball is a complicated sport and stats can't be informative because it's just such a complicated sport? And everyone here knows that volleyball is broken up into 25 point sets. Literally everyone knows that. No one is saying anything different. The whole argument is that some people are wrong because volleyball is too complicated for stats and it's played in separate sets to 25?? That's a pretty weak argument if you ask me.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 15:10:40 GMT -5
Stats can be informative. But they can also be misleading. That's all. We just need to be careful with our conclusions based on stats (especially when they are as woefully inadequate as most VB stats are).
And, yes, the fact volleyball is not a running score (other than the sets won and lost) does make it additionally difficult to analyze. Tennis would be another one.
Match was tied 9-9 in the 5th. I believe -- and maybe I'm wrong -- that this means it could easily have gone either way and that the 10 points that decided it are not going to tell you a hell of a lot because there is going to be a lot of random noise in there.
Is it fair to say that Stanford's serving stats were a major reason why they managed to play Nebraska to 9-9 in the 5th? Maybe. I'm not even sure I'd want to go that far. But maybe. The serving stats had nothing to do with those last 10 points, however.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 15:17:04 GMT -5
Personally, I think something that will tell you a lot more -- and is missing from the analysis above (I think) -- is how the teams performed in each rotation. And how those rotations aligned throughout the match.
My guess (and no I haven't done the analysis) is that Nebraska lost the match when they could not easily get the ball to Foecke or Stivrins (Nebraska has some *really* weak rotations). And my educated guess is that the major reason Stanford won was because Holly Campbell had the match of her career. I tend to look for things that were unexpected.
But, sure, maybe it was the aces.
|
|
|
Post by donut on Jan 27, 2019 15:40:52 GMT -5
is that VB IS an i9diosynchratic sport: broken up into 25-point sets. So this is what you're saying? It's as simple as this? Volleyball is a complicated sport and stats can't be informative because it's just such a complicated sport? And everyone here knows that volleyball is broken up into 25 point sets. Literally everyone knows that. No one is saying anything different. The whole argument is that some people are wrong because volleyball is too complicated for stats and it's played in separate sets to 25?? That's a pretty weak argument if you ask me. I don't think anyone is arguing that because volleyball is a "complicated sport" or consists of 25-point sets (best 3 of 5), that stats are useless. Posters are arguing that a "total points" analysis loses validity in determining a statistical reason for losing a match because "total points" has never determined a volleyball match winner. It's a significant qualification to the claim being discussed. Looking at the data that was provided by Joe, I think the "safe," high-level takeaway is that Stanford won because in 2 sets they won the non-rally game, and in 1 set they won the rally game. That is different than saying they won because they had 7 more no-rally points than Nebraska - the "no rally explanation" doesn't hold for one of their set victories, which again, is how the victor of a volleyball match is determined and thus is significant. Now, that is only with the data presented - @ruffda and @allamerican11 have discussed other buckets of data (currently not captured data around rally events and OOS vs. in-system) that may also offer interesting insights. But just because more and more factors are relevant and posters are identifying assumptions and weaknesses with certain models/approaches, doesn't mean "volleyball is too complicated for stats" - in fact, I think it proves the opposite: that data analysis is imperative for a deep understanding of the sport. BTW, I'm sure Coach Cook's main takeaway from this match was the service line issues Nebraska had. And that's what he will focus on because it is something he can control. So for a coach, maybe the "no-rally game" discussion is a sufficient interpretation of the data, but from a pure data analyst perspective taking into account the structure of the game of volleyball, it isn't a sound interpretation.
|
|
|
Post by Scipio Aemilianus on Jan 27, 2019 16:01:04 GMT -5
Is it fair to say that Stanford's serving stats were a major reason why they managed to play Nebraska to 9-9 in the 5th? Maybe. I'm not even sure I'd want to go that far. But maybe. The serving stats had nothing to do with those last 10 points, however. Thank you for this post. And especially this last line here. Yes you are correct that the serving stats had nothing to do with the last 10 points, and I think that's because the serving stats were fairly even in the last 10 points. Stanford had an ace to make it go from 13-10 to 14-10 If this match (or each set) ended with the exact same number of serving aces and errors for both teams, then no one would be making that argument. The game got to 9-9. But it's not just about how it ended from there but also how it got there. Nebraska had 7 kills and 3 errors and Stanford only had 4 kills with 2 errors. Usually that results in a 2 point lead for Nebraska but their 2 serving errors made it 9-9. Nebraska played better in the first 16 rallies 9-7 but did not have a lead because of the 2 "non-rally points" that Stanford got. Stanford won the next 8 rallies 5-3 (including an ace) to make it 14-12. That's what people seem focused on. If you are going to say those 8 rallies are why Stanford won, why are people also ignoring the previous rallies that made it 9-9 before those 8 rallies? Nebraska easily could of been up 10-8 or 11-7 if it had made their serves and gotten a stop (something they did 4 times in set 5). If that was the case, Stanford's play after the first 18 rallies would not have been enough to win.
|
|
|
Post by Scipio Aemilianus on Jan 27, 2019 16:07:07 GMT -5
I don't think anyone is arguing that because volleyball is a "complicated sport" or consists of 25-point sets (best 3 of 5), that stats are useless. Posters are arguing that a "total points" analysis loses validity in determining a statistical reason for losing a match because "total points" has never determined a volleyball match winner. It's a significant qualification to the claim being discussed. And literally no is looking at "total points" won analysis over the course of the match. We are looking at certain stats in certain specific sets. The total points literally say it's a one point difference. And in volleyball you can't win by one point. So no one is arguing total points over the match is why Stanford won. I don't get this argument that we are clumping match stats as a whole.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 16:12:43 GMT -5
is that VB IS an i9diosynchratic sport: broken up into 25-point sets. The whole argument is that some people are wrong because volleyball is too complicated for stats and it's played in separate sets to 25?? That's a pretty weak argument if you ask me. Agreed. Probably why no one is making that argument. We can debate without purposefully exaggerating and mischaracterizing each other's positions.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 16:18:29 GMT -5
in the third— in their “blowout.” In your honest opinion, what is your reasoning for why Stanford won and Nebraska lost? That's the thing: I'm not claiming there was a singular reason. You and Mike are. I'm comfortable with there NOT being a single stat to point us toward why someone won or lost. The only thing I know for sure is that Nebraska and Stanford were tied and 9s in the fifth and were both in the same position to win the match. What happened after that point (Nebraska's inability to side out, block, dig, transition out, etc.) was significant because it was about executing when the match was on the line. Stanford did that. There's no way to statistically quantify execution under extreme pressure. But Stanford was able to overcome that stress despite their best player having a relatively poor match up until that point.
|
|
|
Post by donut on Jan 27, 2019 16:32:30 GMT -5
I don't think anyone is arguing that because volleyball is a "complicated sport" or consists of 25-point sets (best 3 of 5), that stats are useless. Posters are arguing that a "total points" analysis loses validity in determining a statistical reason for losing a match because "total points" has never determined a volleyball match winner. It's a significant qualification to the claim being discussed. And literally no is looking at "total points" won analysis over the course of the match. We are looking at certain stats in certain specific sets. The total points literally say it's a one point difference. And in volleyball you can't win by one point. So no one is arguing total points over the match is why Stanford won. I don't get this argument that we are clumping match stats as a whole. You misunderstood me. When I say "total points" analysis, I mean looking at total points of non-rally vs. rally points (how Joe originally broke down the box score), and roughly concluding the following: "Nebraska lost because although they had 6 more rally points than Stanford, Stanford had 7 more non-rally points, which means Nebraska lost because of service errors and aces." That general claim was made multiple times in this post alone. I think bluepenquin hit it on the head - there's a major difference between "why did Stanford score more points than Nebraska?" and "why did Nebraska lose?" I think the +7 non-rally points for Stanford versus the +6 rally points for Nebraska sufficiently answers the first question. It doesn't answer the second.
|
|
|
Post by Scipio Aemilianus on Jan 27, 2019 16:53:13 GMT -5
And literally no is looking at "total points" won analysis over the course of the match. We are looking at certain stats in certain specific sets. The total points literally say it's a one point difference. And in volleyball you can't win by one point. So no one is arguing total points over the match is why Stanford won. I don't get this argument that we are clumping match stats as a whole. You misunderstood me. When I say "total points" analysis, I mean looking at total points of non-rally vs. rally points (how Joe originally broke down the box score), and roughly concluding the following: "Nebraska lost because although they had 6 more rally points than Stanford, Stanford had 7 more non-rally points, which means Nebraska lost because of service errors and aces." That general claim was made multiple times in this post alone. I think bluepenquin hit it on the head - there's a major difference between "why did Stanford score more points than Nebraska?" and "why did Nebraska lose?" I think the +7 non-rally points for Stanford versus the +6 rally points for Nebraska sufficiently answers the first question. It doesn't answer the second. I apologize to you if I misunderstood what you wrote. Or if I didn’t clarify something. Nebraska being +6 and -7 in rally/non-rally points is a match total. That means nothing to me. That match total is irrelevant. Stanford could of won 25-23, 25-23, 10-25, and 25-23 and they would have lost in that match point total stat by a lot. Which is why that doesn’t matter to me. What matters is by set. In this match, I only care to know why Stanford was +2 in set 1 and +3 in set 5. Even why Nebraska was +3 in set 2. Why did Nebraska lose set 1? Maybe because Stanford had 5 aces and 2 errors while Nebraska had 1 ace and 2 errors. If Nebraska wins set 1, we can assume they would win in 4. Same type of thing can be done for set 5. Starting at 0-0, not some random tied score while forgetting what happened before to get to that point.
|
|
|
Post by donut on Jan 27, 2019 17:12:14 GMT -5
You misunderstood me. When I say "total points" analysis, I mean looking at total points of non-rally vs. rally points (how Joe originally broke down the box score), and roughly concluding the following: "Nebraska lost because although they had 6 more rally points than Stanford, Stanford had 7 more non-rally points, which means Nebraska lost because of service errors and aces." That general claim was made multiple times in this post alone. I think bluepenquin hit it on the head - there's a major difference between "why did Stanford score more points than Nebraska?" and "why did Nebraska lose?" I think the +7 non-rally points for Stanford versus the +6 rally points for Nebraska sufficiently answers the first question. It doesn't answer the second. I apologize to you if I misunderstood what you wrote. Or if I didn’t clarify something. Nebraska being +6 and -7 in rally/non-rally points is a match total. That means nothing to me. That match total is irrelevant. Stanford could of won 25-23, 25-23, 10-25, and 25-23 and they would have lost in that match point total stat by a lot. Which is why that doesn’t matter to me. Awesome, we agree! I'm not a fan of saying if Team 1 won set X they would have won the match, because sets don't happen at the same time. They happen in a linear fashion, where set 2 is often a reaction to set 1. What is the difference between starting at 0-0 and starting at 9-9 though? At both points in the set, all things are equal. Both teams have to get to 15 (or 6, I guess, in the case of 9-9), first? In theory, it doesn't really matter how they got to that point - the slate gets wiped clean.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2019 17:31:54 GMT -5
You misunderstood me. When I say "total points" analysis, I mean looking at total points of non-rally vs. rally points (how Joe originally broke down the box score), and roughly concluding the following: "Nebraska lost because although they had 6 more rally points than Stanford, Stanford had 7 more non-rally points, which means Nebraska lost because of service errors and aces." That general claim was made multiple times in this post alone. I think bluepenquin hit it on the head - there's a major difference between "why did Stanford score more points than Nebraska?" and "why did Nebraska lose?" I think the +7 non-rally points for Stanford versus the +6 rally points for Nebraska sufficiently answers the first question. It doesn't answer the second. Why did Nebraska lose set 1? Maybe because Stanford had 5 aces and 2 errors while Nebraska had 1 ace and 2 errors. I mean, no... All five of Stanford aces happened at or before going up 21-19 in the first. Their five aces in the first set were almost wiped clean by Nebraska because Nebraska FBSO'd Stanford's next four serves and had a good run on Miller's serve, blocking/digging well enough to get the ball to Stivrins in trans, for three straight points. Both teams were tied at 26s, and then Sun made two errors to end the set. Just like you can't look at matches as a whole, you also can't view sets that way. This set included 54 total points; 5 of those points being aces is meaningless unless you consider WHEN in those 54 points the aces happened. In this case, the aces were not significant in deciding the set. Nebraska nearly won the set despite being aced five times.
|
|