Recent Entries
 | |
 | Do the Strong always win? Posted by |
This season there seemed to be quite a few comments on matches noting various points on strength of opponent, home v away, etc. that all pointed to a general assumption that the perceived stronger team is going to win. Generally, that's true. An equally important point is to note how much you can increase the odds for a win, or at least winning 1-2 sets and gathering the XP and CP, by closing the skills gap.
Using the Ratings, as those are availble in all match results and are the closest thing we have to a relative strength stat, here is how Season 9 played out in answering the question: How often does the Stronger Team win?
There were 537 total matches. Of those, 18 had identical ratings, the equivalent of an 'evenly matched opponents.' We'll exclude those, because nobody was stronger.
In the 519 matches, the stronger team won 439 (84.6%). Of those, 274 were in 3 sets, 106 were 4, and 59 won in 5. Of the losses, 17 were straight sets, 30 were 4 set losses, and 33 were 5 set losses. So a straight 3-set loss by the stronger team was rare (3.3%). What skews those numbers so high?
In matches where one team was 13 or more points higher per player, those teams went 179-0. 159 were 3-set wins, and only 2 matches went to 5 sets. So when there is a large skill gap, the current program definitely tends to reward the stronger team and the results bear that out.
So what happens when the skill gap closes?
Stronger team 10-12 points higher: 56-4 (93%). 6 to 9: 77-14 (85%). 2 to 5: 102-44 (70%). 1 pt: 25-18 (58%). Additionally the percentage of 3-set wins decreases and the number of 5-set wins increases as the power levels even out.
How much does Home Court play in?
Home teams were 256-25 (91%) when stronger. Again, teams 13+ were perfect at 131-0. 10-12: 35-2 (95%). 6-9: 34-5 (87%). 2-5: 51-11 (82%). 1 pt: 5-7 (42%).
Meanwhile, Away teams that were stronger did not win as frequently. Overall, away teams were 62-25 (71%) when stronger than the home team. 13+: 11-0 (100%). 10-12: 10-2 (83%). 6-9: 19-3 (86%). 2-5: 14-14 (50%). 1 pt: 8-6 (57%).
On neutral courts, the stronger teams were 121-30 (80%). 13+: 37-0. 10-12: 11-0. 6-9: 24-6 (80%). 2-5: 37-19 (66%). 1 pt: 12-5 (71%).
Smaller sample sizes always lend lots of variation to statistics, so take all the above as just one season and not definitive of the program overall. What can we take away from this though?
Fluke results where a clearly overmatched team steals a victory outright are rare, and greatly taper off at a 10-pt skill difference, disappearing over 12. If you can close the gap on a top team, even on the road, to less than 10, you'll have at least a 10% chance of winning. Get it to within 5 points, and that jumps to 30%. And remember that's an average of 5 points per player! Not total skills points combined. If 3-4 of your team are solid, having 2-3 weaker players is not an automatic kiss of death (unlike Al asking if you'll go undefeated next session). So if your T7 can't quite match up, you can still push for extra sets and hope they rest a top player or split their squad out to close the gap even further. Even one of their T7 being out can drop their average 3 pts pr more. To give even more hope to the underdogs, home teams that were just a point higher (that's after including the home bonus) had a losing record this season.
Other thoughts or insights? Or data points missing from above that you'd like to see? I didn't throw out the whole table, yet I can drill down to individual teams if anyone wanted to know how their own team fared Home/Away/Neutral on actual vs "expected" results.
Readers Comments
This is brilliant, Kev. I was hoping to do something similar but the computer packed up and trying to do this on the phone is near on impossible.
Of course I'm curious I we were as unlucky as it felt at times but we aren't exactly a strong team. Maybe a simple count of how often we were the stronger (6+]., weaker (also by 6+) if close (-5 to +5).
Many thanks, great and valuable info.
I was thinking of doing some calculations around these numbers myself, but your aggregate numbers are good enough for my concerns.
On a sidenote:
Quite probably your "bad" finish to the season (only 3 trophies) was based on your 1st round cup match strategy.
By planning a Shield run for NO (via losing to SB) you showed everybody else that your squad can be defeated.
Which meant effectively more T7 (instead of B7) lineups vs NO, ergo more expectable losses.
Would you mind electronically mailing me the data, Kevin?
Somehow, I think this article is aimed at me....
Superb analysis Kevin.
I'd love to see the data for Pekin, but it would also be great to see actual vs expected for all teams just to know who Olmec really loves.
Steve, CCT's tallies:
When CCT was +8 or better, you went 8-0 (6-0 Home, 2-0 Neutral). 5 were 3 sets sweeps, 3 were 3-1 win. Avg 6.5 pts/set better.
+5-7, CCT went 2-3 (1-1 H, 0-1 A, 1-1 N). You got outscored by a grand total of 9 pts over 20 sets.
+2-4, CCT was 2-1 (1-1 H, 1-0 A). 2 went 5 sets, 1 went 4. Total pts CCT was outscored by just 2 pts over 14 sets.
+1, 0 or -1. This is where your luck went south. When really close, CCT went 0-6. (0-2 at home, both losses were 3 set sweeps by the Scarecrows and they outscored you by 36 pts in 6 sets. 0-1 A, 4 set loss. 0-3 neutral, 1 3 set, 2 4 sets, -35 pts over 11 sets.
-2 - -4. 0-1 A (5 sets, yet outscored by 16 pts total)
-5 - -7. A bit lucky here. 2-2 total (1-2 A, 1-0 N). Close matches, with CCT outscoring the opponents by just 0.5 pts/set total.
-8 or less. 0-5 (0-4 A, 0-1 N). 3 were sweeps, 1 5-set. Outscored total average of 4 pts/set.
Splitting the -1, 0, 1 matches would have landed CCT a winning season at 17-15 (or 17-16 if advancing in a cup and losing the next match). Going 6-7 in the 2-7 range (+ or -) balanced out, and CCT avoided any of the massive upsets either way at 8 pts or higher.
Thanks, Kevin, that's great. Three points:
1. I mention in some previews that a team is 5 points a man better "but that means nothing". Seems I was right there from my perspective even if over all it's more like 80-20
2. 0-6 when -1 to +1 wow I think we can agree Olmec hated us.
3. Maybe by player positions were all wrong or maybe it shows what a weak setter and/or right side can do to your team.