with all these talk on service vs. rally scoring system. i think it is worthwhile to look into the core of the debate, ie. the badminton service, and what role does it play in a badminton game. from the purely statistical point of view, i have taken all the recorded match of the 2005 World Championship, and derived some statistics about the service. i said recorded matches because i was only able to obtain data from 234 matches in the WC 2005. some matches' data are missing. here are the statistics that i think are interesting: overall conversion rate : 47.8% ie. if a player has the service, the probability that he wins that rally. winner conversion rate : 54.4% ie. the conversion rate for the match's winner loser conversion rate : 39.3% ie. the conversion rate for the loser of the match. the statistics is interesting. firstly, it tells us that, overall, it will take a player 2 services in order to score 1 point. this illustrates the neutral nature of the badminton service. ie. the side who has the service does not hold an advantage in the rally. this is similar to volleyball. but the opposite of tennis, where the server in general has a definite advantage to win the rally. thoughts? comments? conclusion? i have my own, but i think it will be interesting to hear what other think first given the above data.
no. as i mentioned, it was extracted from the 2005 World Championship. where service base scoring was used. it will be interesting to see if a rally scoring system significantly changes the conversion rate. but i don't have any access to those data.
Sorry kwun, a bit disorientated. But yea you still need the rally system to compare the impact. If this stats are correct then I think we would see tighter games as as you mentioned 1/2 serves is a point hence the other player would have also have a point meaning that we might see either very very close games or just plain ol fashioned demolition
where have the figures come from? TV network, IBF, you watched the games and recorded the data yourself? I ask because the numbers feel wrong. I'm not sure why yet, it's just a feeling.
i went to tournamentsoftware.com. in the WC 2005 results, they have data on the number of services for each match/game. the percentage was calculated using that data and the final score of the match. the final value was calculated using all 234 matches they have data on. so it is quite sound statistically as the sampling population is not that small. however, i am not a statistician so i cannot say how accurate it really is.
I have no doubt that this will change with rally scoring. Simply because the server risks to lose the point, while the situation for the receiver is unchanged. In other words, "playing safe" when serving will become more important. Taking risks in the serve, such as aiming to (almost or slightly) touch the the net tape and hit the lines, will not become a hopeless strategy. This will leave room for even more aggressive returns. Thus, the above conversion rate will decrease -- and the power balance is lost. I suspect the percentage of high serves will rise, thus making the serve an even more defensive shot than it is today. As I've said before, I see this as the biggest problem with the new scoring system, because this really changes the nature of the game. IHMO, the other changes are superficial in comparison.
Are the figures for all categories, singles and doubles? If they are, I think there should be seperate figures. I suspect the conversion rate for doubles will be much less (especially for MD, where the receiver almost looms over the net).
When was the serve ever an offensive weapon? It can only be offensive if the receiver of serve is overly offensive (eg. rushing to the net), but by itself, a serve can never be offensive. Edit: (Of course, I am not about weak players who may have problems retrieving drive serves.)
Another thing I am puzzled about. Sorry if I am wrong. Shouldn't the overall conversion rate be (winner rate + loser rate)/2 ? Edit: Oh, I realise I am wrong. Because the loser serves less number of times, so the overall figure can be skewed more towards the winner's rate. eg. Winner serves 100 times. Rate 56%. Loser serves 50 times. Rate 36% So, overall rate = (56+18)/150 x 100 = 49.333%
ops. you're right there. i have incorrectly weighted the result on a per game basis. i will fixed the values in 1 min.
actually, i take that back. if you look at the 4 data i used: winner points loser points winner services loser services overall conv = (winner + loser points) / (winner + loser services) winner conv = winner points / winner services loser conv = loser points / loser services mathematically, they do not add up. but either way, i have fixed the percentage, now i am calculating using the totally sum of points/services instead of calculating the conversion rate per match and then averaging.
unfortunately, data for the rally scoring system isn't available yet. it will be very very interesting to see the equivalent statistics. most ppl do agree with your prediction. however, i myself is still unsure if that is indeed true. i was watching the World Cup mixed final. i didn't pay much attention to the socring but instead the rallies themselves. and frankly i had a hard time trying to sense any difference in the technical nature of the rallies themselves. the mental pressure of the players, as well as the tension and excitement audience senses, on the other hand, will be difficult to derive using number alone. so while this is still an worthy exercise, tension and excitement are the qualities that us fans look for in a match.
I was just having a look at the website, and I was about to ask you if you "winner" was the overall match winner or game winner I would have thought that by game would be more meaningful
i gave it a few minutes of thought, and i am still not sure which is more correct way to do it. can you give your reasoning?
and I think some of the data is a little incorrect e.g. second match I looked at Kennevic Ascuncion v Andrew Smith game1: KA won 17-14 with 25 serves, loser had 24 serves you can't get 3 points ahead with only 1 serve more than your opponent game2: AS won 15-5 with 20 serves, loser had 12 serves we know KA served first because he won the first game so how does AS get 10 points ahead with only 8 serves more than KA I think the difference in points can only be 0 or 1 different from the difference in number of serves. 0 if the winner of the game served first 1 if the loser of the game served first (unless there's a red card, I suppose, but that's very very rare) game3 looks OK game1: KA won 17-15 with 30 serves, loser had 29 serves 2 point diff, 1 serve diff, but we know the loser served first so we expect a difference of 1 between the point diff and serve diff
Some sloppy reading there on your behalf I'm afraid. I didn't say that the serve today is an offensive shot. In badminton, as opposed to most other rally sports, the very nature of the service is defensive. I said that it will become even more so with rally scoring. And I forgot to add that I was talking about doubles. I don't think the rally scoring will radically change the way that singles is played.
err not sure. it was another feeling. but by match or by game will give different results for winner and loser, but should give the same result for overall. that game KA v AS was 17-14 5-15 17-15 by game you get winner 65% loser 52% overall 59% by match you get winner 58% loser 60% overall 59% the match winner won less points and had less serves than the loser if you do it by match, you have no idea how much the figure is affected by matches going to 3 sets
