Disclaimer: CurlingZone does not represent or endorse the accuracy or reliability of any Content posted on any Forums area and you acknowledge that any reliance upon such Content shall be at your sole risk. Any Content placed on any Forums area by users and anonymous posters are the views of the user posting the statement, and do not represent the views of CurlingZone or our partners, advertisers or sponsors. By posting anonymously, you are allowing your IP address to be displayed for identification purposes. CurlingZone reserves the right to remove any post at its discretion without warning or explanation.
04-04-17 11:04AM |
|
curlky
Drawmaster
Registered: Oct 2013
Location:
Posts: 559 |
The Curling Percentage System Is Broken
So I watched the USA v Canada matchup last night, and saw how the US, specifically Shuster, had some bad shots at bad times that led to a beat down. After watching it, and then looking at the post game box scores, I feel like that game is specific proof that the curling percentage system is completely broken and needs to be fixed.
(you can skip this paragraph if you watched the match) Good example, 1st end, down to rocks 15 & 16 of the end, Shuster does not have hammer. Gushue is sitting 2, with shot rock being left side of button at tee line, sort of buried so no chance to remove it, and then a rock in the back right 12 foot is the second rock. Shuster can hit the second rock out, and then Gushue has a pretty easy draw for 2. Shuster opts to play the corner freeze onto shot rock, where if he makes it, he forces Gushue to 1 most likely, or forces Gushue to throw a sort of tough hit and roll to take two. Shuster throws his shot a bit too heavy, and ends up on the edge of the back 4 foot, completey unguarded. This leaves Gushue a simple nose hit to take 3.
(OK now to my point) Shuster’s throw was scored a 50% on this throw. How is this possible? I cant think of a worse outcome for his shot. If he is more heavy, they can sweep it back, and maybe it buries ni behind the shot rock, and Gushue has at least a hard shot for 3. If he is 4-6 feet shorter, it makes the hit a bit harder as at least it sets up a small risk of the jam from Gushue’s last rock.
How can this shot be scored 2 out of 4? What would have scored a 1 or a 2? I realize that the scoring does nto take into account pure strategy to know if a shot was a wise call or not, but in this case, the shot clearly led to an extra point by the other team, and should have been a 1 or a 0 in my book.
And then we can look at the 6th end, going into Shusters last rock. Canada sits 3 or 4, and Shuster looks like he is going to be forced to take 1. All he has to do is throw a light weight nose hit to an unguarded rock in the 4 foot with a slight roll towards to center (1 inch to 2 foot would have been OK) and Shuster takes 1. If he just nose hits and sticks, or nose hits and roll out 3 to 4 foot, he gives up 1. Instead he overcurls but still manages to hit the shot rock, but rolls out and gives up 3. This shot is scored a 1 out of 4. I guess a 0 out of 4 would have been a complete miss, but I don’t understand why this shot is worthy of scoring any points at all.
So long story short, I don’t mean this to comment on how well or poorly Shuster played, but more of how the scoring system is completely flawed. It is my belief that just talking about those 2 ends, good play would have ended up with Gushue taking 1, and Shuster taking 1. Slightly worse play would have led to Gushue taking 1, and then Gushue taking 1 more. Poor play would have led to Gushue taking 2 and then Gushue taking 2 more. Really bad play would have led to Gushue taking 3, and then taking 3, and hogging both rocks would lead to Gushue taking 3 and then 4.
So to summarize
Good: 1-1
OK: 2-0
Bad: 4-0
Awful: 6-0
Hogging both: 7-0
So the game outcome was 6-0, and Shuster scored 3 out of 8. Out of all the outcomes, does a score of 3/8 seem right? To me, it does not. I would propose a solution like this. The scoring system allows that rocks be scored 0 to 4, but bonus points can be scored for really hard shots. I think that really awful shots be scored lower than 0, specifically for skips last rocks when points can be directly attributed to outcomes. Under that system, I would score Shusters rock in the first end as 0 out of 4, maybe 1 out of 4. Then I would score his last rock in the 6th as a -1 out of 4. This would give him a score of 0 out of 8, or -1 out of 8.
You might say this is too harsh, but realistically, I think that if you went up to John and asked him to score his 2 shots, he would not give himself 3 out of 8.
Thoughts?
Report this post to a moderator | IP: Logged
|
|
04-04-17 11:50AM |
|
Three
Swing Artist
Registered: Feb 2015
Location: Canada
Posts: 278 |
Of course you are right. I take curling percentages with a 22lbbsg of salt. Way too many variables. When curlers debrief after a game with their coach they look at makes and misses and what was the correct miss for any situation, etc,...not useless percentages. Trying to put a number on it is especially dumb when done by a third party. If a perfect throws is overswept by a bad line call what is the score for that stone? Curling is a team game, this is not baseball.
Report this post to a moderator | IP: Logged
|
|
04-04-17 12:10PM |
|
broomsmith
Harvey Hacksmasher
Registered: Apr 2016
Location:
Posts: 13 |
Percentages are great tools for leads.
Percentages provide minimal information value for skips.
Report this post to a moderator | IP: Logged
|
|
04-04-17 12:22PM |
|
prairie guy
Swing Artist
Registered: Jan 2011
Location: Southern Sask.
Posts: 364 |
Yes percentage comparisons are OK for front end players but not necessarily for the back end. Some people get worked up about the accuracy of percentages (especially one guy I curl with he thinks if the skip has to draw to the button against 4 but ends up with a piece of the 4 foot and gives up a steal of 2, this should be scored as zero points!)
Anyway, when TSN gives their percentage updates during the game I don't let that influence how I feel the players are shooting. Percentages perhaps can be utilized to measure the performance over the long term but even that requires a fair amount of subjectivity.
Report this post to a moderator | IP: Logged
|
|
04-04-17 12:39PM |
|
TNH
Hitting Paint
Registered: May 2011
Location:
Posts: 161 |
The percentages are for the fans, not the players. What's going on with this game - who's on fire and who can't buy a shot?
The other thing is that stats get better with more data. So a percentage on an individual shot is kind of meaningless, but they do tell a story over the course of an event.
So with that in mind, and as a fan I don't really mind them. Adds a bit of emotion when following the games CurlingGeek-style on world curling.org.
Report this post to a moderator | IP: Logged
|
|
04-04-17 02:09PM |
|
curlky
Drawmaster
Registered: Oct 2013
Location:
Posts: 559 |
I think that my point is missed here. I understand that stats can be misleading, due to the simple fact that the scorer does not always understand that call, and does not account for poor strategy.
My point is that within the system that is being used, how did Shuster receive the scores that he did. How does the system allow this? And if it does allow this, arent we better off without any system that the one we have?
Report this post to a moderator | IP: Logged
|
|
04-04-17 03:13PM |
|
TNH
Hitting Paint
Registered: May 2011
Location:
Posts: 161 |
Ah, I see. I do tend to agree that questionable values from individual shots can also make the overall percentages more questionable.
I can't answer your question though. Even during US-Canada I would think "oh oh, that stone's getting hit soon" and Kevin Martin would say "All things considered, pretty good shot!".
I did scan a couple of scoring guidelines (like: www.curling.ca/wp-content/uploads/2...ng-v2013.1.pptx) and two things struck me. The first was "when in doubt, favor the player" which makes sense to me.
The second was the 0-4 scale for "did the shot improve the position". In your Shuster example, the freeze bounced open but was still shot stone - Gushue still had to make a play. So I agree the short wasn't made, but by the guidelines he did improve his position and I can see the partial score.
Making it better is a tough problem.
Report this post to a moderator | IP: Logged
|
|
04-04-17 04:25PM |
|
dbsdbs
Drawmaster
Registered: Feb 2013
Location:
Posts: 812 |
quote: Originally posted by TNH
The percentages are for the fans, not the players. What's going on with this game - who's on fire and who can't buy a shot?
That is it exactly. Just like baseball fans like ERA and BA stats, even though there are so many better measurements of success. Curling percentages give the fans a relative comparison among players/teams and give the announcers something to talk about, and that is all they do.
Report this post to a moderator | IP: Logged
|
|
04-04-17 06:29PM |
|
curlky
Drawmaster
Registered: Oct 2013
Location:
Posts: 559 |
quote: Originally posted by dbsdbs
That is it exactly. Just like baseball fans like ERA and BA stats, even though there are so many better measurements of success. Curling percentages give the fans a relative comparison among players/teams and give the announcers something to talk about, and that is all they do.
There may be better stats in baseball than ERA or BA (such as WHIP or WAR) but at least they are stats that have a basis for accuracy. As it stands now, especially for a skip, the percentage is neither helpful or accurate. We need to band together and not stand to keep such a silly stat. We need to get something better in place. Or at least tweek what we have now.
Report this post to a moderator | IP: Logged
|
|
04-04-17 07:56PM |
|
IN-OFF-FOR-2
Super Rockchucker
Registered: Mar 2013
Location:
Posts: 1875 |
quote: Originally posted by curlky
I think that my point is missed here. I understand that stats can be misleading, due to the simple fact that the scorer does not always understand that call, and does not account for poor strategy.
My point is that within the system that is being used, how did Shuster receive the scores that he did. How does the system allow this? And if it does allow this, arent we better off without any system that the one we have?
I've said this in the past and I'll say it again, there is waaaaaay to much left up to interpretation by the stats keepers. Using Gushue as the example the other day it was stated he curled 100%. Impossible. He missed at least 3 shots including one that was supposed to be a freeze but instead ticked a guard and rolled over. Take whatever stats that are declared and reduce by 20% for everyone.
Report this post to a moderator | IP: Logged
|
|
04-04-17 07:58PM |
|
Grat
Hitting Paint
Registered: Mar 2014
Location:
Posts: 107 |
Good luck coming up with a better stat. Just like batting average won't tell if a ball is hit hard right at someone, or it's a strike out with two on in the bottom of the ninth, curling percentages, figured objectively, are just going to tell you how close shots are to making the called shot.
Especially in a single game with a small sample size the percentage isn't always the best way to measure performance. But an attempt to make it more meaningful would also make it more subjective, and even more prone to error and odd looking results.
Report this post to a moderator | IP: Logged
|
|
04-05-17 12:25AM |
|
victormlee
Knee-Slider
Registered: Mar 2017
Location: Massachusetts, USA
Posts: 2 |
quote: Originally posted by curlky
I think that my point is missed here. I understand that stats can be misleading, due to the simple fact that the scorer does not always understand that call, and does not account for poor strategy.
My point is that within the system that is being used, how did Shuster receive the scores that he did. How does the system allow this? And if it does allow this, arent we better off without any system that the one we have?
Take a look at http://www.curlit.com/powerpoint/HowToScore-Engl.pdf. This is the system used for WCF events.
The scorer merely considers the intended call and the outcome. Strategy has nothing to do with the scoring. This is all computerized for the WCF top events. Page 10, lower half of the PDF link I gave probably explains why the scorer gave it as a 2/4. It was a draw. The draw landed in the house, but in the wrong spot.
However, I am a little confused by page 15. If, as you say, the call was a freeze, then it would seem a miss that was that heavy (more than 4" away) should have been scored 0/4.
You can legitimately debate the merits of the system. I think what you are really arguing for is a scoring system that is better correlated with outcomes.
Last edited by victormlee on 04-05-17 at 12:39AM
Report this post to a moderator | IP: Logged
|
|
04-05-17 09:49AM |
|
Clappy
Harvey Hacksmasher
Registered: Feb 2017
Location: Westland, MI
Posts: 70 |
I'm not sure how your second example wasn't scored zero. I thought I had read in the rules, that for a last rock, the scores were either 100% or 0%, basically a pass/fail based on the task being accomplished regardless of how well or poorly it was executed.
Report this post to a moderator | IP: Logged
|
|
04-05-17 09:58AM |
|
jhcurl
Super Rockchucker
Registered: Sep 2002
Location: US - CT
Posts: 1431 |
The other thing to understand about scoring is the scorer does not have the benefit of listening to the players. Having scored at Nationals and Worlds that sometimes makes it difficult to know the intended shot and if it was made. A shot after a timeout is the worst. Players pointing at multiple options, then the broom goes down. Usually have to watch the shot to figure out what was intended.
JH
Report this post to a moderator | IP: Logged
|
|
04-05-17 01:31PM |
|
Itsjustagame
Hitting Paint
Registered: Aug 2014
Location:
Posts: 106 |
Re: The Curling Percentage System Is Broken
quote: Originally posted by curlky
I feel like that game is specific proof that the curling percentage system is completely broken and needs to be fixed.
When I have something broken and useless, I simply throw it away.
Report this post to a moderator | IP: Logged
|
|
|
All times are GMT. The time now is . |
|
|
|
|
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is OFF
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|