View Full Version : Fouls and Tiebreakers
There's been a lot of talk about the final Einstein match, where the win was swept up by the 1st Tiebreaker level: foul points. As a lover of stats, I decided to run an analysis and see how common this was throughout the season. I've split foul-decided matches into two categories:
The first type (Tie) is a match where the ending scores are tied, and the 1st tiebreaker is used to determine the winner (i.e. both alliances score the same, but the alliance creating the least fouls wins)
The second type (Sway) describes a match where fouls 'sway' the final score of the match (i.e. red scores 100, blue scores 97, but red got 5 points due to a blue foul. If this foul didn't occur, blue would have won, hence 'swaying' the match)
Let's take a look, starting off at the Top.
Einstein Field
Ties: 1 (Finals 1 Match 3)
Sways: 3 (Quarters 1 Match 1, Semis 1 Match 1, Semis 1 Match 2)
Total: 4 foul-decided matches
Championship Divisions
Ties: 8 (Arch: 2, Carson: 1, Curie: 3, Galileo: 1, Hopper: 1)
Sways: 46 (Arch: 9, Carson: 5, Carver: 8, Curie: 3, Galileo: 5, Hopper: 2, Newton: 9, Tesla: 5)
Total: 54 foul-decided matches
Season Events
Ties: 87
Sways: 587
Total: 674 foul-decided matches
Thoughts?
To give an idea of perspective, there were a total of 13302 matches this season, meaning 5.06% of all matches this year were decided by foul points (0.65% as Tiebreakers, 4.4% as Sways)
You might want to include Red Cards as their own category--that's a Sway, only more devastating. I'm not sure they currently are counted.
Easiest way to find 'em is to find elims matches where an alliance scored zero points. That alliance got a red card and thus lost the match.
The Disable fouls (usually, somebody's bumper fell off) decided some playoff matches, too, but nowhere near as many as red cards.
You might want to include Red Cards as their own category--that's a Sway, only more devastating. I'm not sure they currently are counted.
Easiest way to find 'em is to find elims matches where an alliance scored zero points. That alliance got a red card and thus lost the match.
The Disable fouls (usually, somebody's bumper fell off) decided some playoff matches, too, but nowhere near as many as red cards.
Red cards aren't currently counted.
The easiest way to get them is actually to take the total points, add the adjustments (negative for red cards) and check if it's zero. If I recall, there was a match this season where an alliance scored 0 points without a red card or any penalties.
If you include red cards only, there were 18 matches where the alliance that got the red card would have won if not for the red card. 4 of those were in Champs Divisions (Newton, Carson, Carver, Galileo with 1 each), and 0 on Einstein.
Unfortunately I can't detect disable fouls, as the match score breakdown doesn't specify what the fouls were for, nor if any robots were disabled in the match.
orangemoore
06-05-2016, 22:40
For the season events is that only for Playoffs?
If not, do you know the number of ties in Playoffs only? It would be interesting to see how many matches would have been replayed if there weren't any tie breakers.
For the season events is that only for Playoffs?
If not, do you know the number of ties in Playoffs only? It would be interesting to see how many matches would have been replayed if there weren't any tie breakers.
It's qualifications and playoffs.
If you include only playoff matches, there were 109 matches decided by fouls, 18 of which were ties, and 91 were sways
Easiest way to find 'em is to find elims matches where an alliance scored zero points.
Here's all the zero-score playoff matches.
It's qualifications and playoffs.
If you include only playoff matches, there were 109 matches decided by fouls, 18 of which were ties, and 91 were sways
In qual matches, ties remain ties. So I do believe the important thing is that 18 playoff matches were decided by tie. (OK, including the final one...)
In qual matches, ties remain ties. So I do believe the important thing is that 18 playoff matches were decided by tie. (OK, including the final one...)
Good call, I missed that part
AndrewPospeshil
07-05-2016, 01:01
It's qualifications and playoffs.
If you include only playoff matches, there were 109 matches decided by fouls, 18 of which were ties, and 91 were sways
That's interesting - if playoff ties were to be decided by a third match, that would be another 18 matches this season, or 45 total minutes of game play. Not that anyone isn't really against it already, but data seems to show that the current playoff tiebreaker rules aren't worth keeping.
That's interesting - if playoff ties were to be decided by a third match, that would be another 18 matches this season, or 45 total minutes of game play. Not that anyone isn't really against it already, but data seems to show that the current playoff tiebreaker rules aren't worth keeping.
I would say that they are worth keeping.
However, that is IF the game doesn't lend itself to ties, which can be extremely hard to predict before the season. 2010 was notorious for its ties (and the low scores compared to most years). 2011-2015, not so much. 2016 was close enough to possibly want it, but I think it was unnecessary. 2017, who knows?
That's interesting - if playoff ties were to be decided by a third match, that would be another 18 matches this season, or 45 total minutes of game play. Not that anyone isn't really against it already, but data seems to show that the current playoff tiebreaker rules aren't worth keeping.
I would say that they are worth keeping.
I think you're both right, for eliminations play a 4th rubber match, AND have a tiebreaker in that 4th match only. (This prevents problems of never-ending eliminations.) But the rules for the tiebreaker have to be non-subjective - ie, not based on fouls.
And in case someone wants to bring up schedule pressure, while a valid point, remember FIRST calls FRC "the ultimate Sport for the Mind." Virtually no sporting events have a hard, fixed schedule that dictates what time the event ends. The game ends, when the game ends. I bet if after Einstein F3 they called a tie and a 4th match the crowd would have gone just as nuts to see that 4th match as they did for the champs being crowned.
Richard Wallace
07-05-2016, 07:23
If you include only playoff matches, there were 109 matches decided by fouls, 18 of which were ties, and 91 were sways
How many equal-score playoff matches were decided by tiebreaker sort order > 1st? (Not by fouls.)
I know of one (http://www.thebluealliance.com/match/2016micmp_ef2m2) that was decided by the 4th order sort. Were any decided by sort order > 4th?
How many equal-score playoff matches were decided by tiebreaker sort order > 1st? (Not by fouls.)
I know of one (http://www.thebluealliance.com/match/2016micmp_ef2m2) that was decided by the 4th order sort. Were any decided by sort order > 4th?
The highest sort order we've reached is 4th. That was achieved by 3 matches:
NE District - UMass-Dartmouth Event - Semis 1 Match 2 (http://www.thebluealliance.com/match/2016manda_sf1m2)
Michigan State Championship - Eights 2 Match 2 (http://www.thebluealliance.com/match/2016micmp_ef2m2)
FIM District - Standish-Sterling Event - Quarters 4 Match 1 (http://www.thebluealliance.com/match/2016mista_qf4m1)
On a slightly unrelated, but still interesting note:
There have been 311 instances where an alliance Tower has ended with a higher strength than what it started with due to Tech Fouls (>8 for Events, >10 for Championships)
Jon Stratis
07-05-2016, 09:17
That's interesting - if playoff ties were to be decided by a third match, that would be another 18 matches this season, or 45 total minutes of game play. Not that anyone isn't really against it already, but data seems to show that the current playoff tiebreaker rules aren't worth keeping.
But for each event a playoff match was decided by a tie, there would be, at minimum, 6.5 minutes added to the schedule, once you include field reset for the typical match turnaround. Add in another 5 minutes for an expected field timeout caused by teams having to play back to back matches. Sure, it doesn't really sound like much - another 11.5 minutes... But multiplied by 18 you have several hours added to event schedules across FIRST. And they are all added on the final day of the event, when teams with a long drive home might be itching for things to get over (especially if that team is no longer playing and is just waiting for the awards) so they can get on their bus.
In other threads people have been exploring the length of Einstein... If we got rid of a tiebreaker, we would have had to add another back to back match, meaning a field timeout, commentary, videos... It would have made the event even longer than it already was.
Kevin Leonard
07-05-2016, 09:25
Here's all the zero-score matches.
You're missing the one in Finger Lakes Quarters at least. http://www.thebluealliance.com/match/2016nyro_qf3m2
Richard Wallace
07-05-2016, 09:39
The highest sort order we've reached is 4th. That was achieved by 3 matches:
NE District - UMass-Dartmouth Event - Semis 1 Match 2 (http://www.thebluealliance.com/match/2016manda_sf1m2)
Michigan State Championship - Eights 2 Match 2 (http://www.thebluealliance.com/match/2016micmp_ef2m2)
FIM District - Standish-Sterling Event - Quarters 4 Match 1 (http://www.thebluealliance.com/match/2016mista_qf4m1)
That last one appears to have been decided by the 5th order sort, cumulative sum of scored tower goal points.
You're missing the one in Finger Lakes Quarters at least.
Thanks. I corrected the table. I found 3 Playoff matches with zero score that had no adjust points.
Thanks. I corrected the table. I found 3 Playoff matches with zero score that had no adjust points.
On the file I downloaded from your previous post, NECMP Final 1 appears twice, with no score to either alliance, where TBA shows 210(blue) - 180(red).
The FRC API has some spurious data in the NECMP "Score Details" endpoint:
{"matchLevel":"Playoff","matchNumber":18,"AudienceGroup":"GroupC","Alliances":[{"alliance":"Blue","robot1Auto":"None","robot2Auto":"None","robot3Auto":"None","autoBouldersLow":0,"autoBouldersHigh":0,"teleopBouldersLow":0,"teleopBouldersHigh":0,"towerFaceA":"None","towerFaceB":"None","towerFaceC":"None","towerEndStrength":0,"teleopTowerCaptured":false,"teleopDefensesBreached":false,"position1crossings":0,"position2":"A_ChevalDeFrise","position2crossings":0,"position3":"C_Drawbridge","position3crossings":0,"position4":"B_Ramparts","position4crossings":0,"position5":"D_RockWall","position5crossings":0,"foulCount":0,"techFoulCount":0,"autoPoints":0,"autoReachPoints":0,"autoCrossingPoints":0,"autoBoulderPoints":0,"teleopPoints":0,"teleopCrossingPoints":0,"teleopBoulderPoints":0,"teleopChallengePoints":0,"teleopScalePoints":0,"breachPoints":0,"capturePoints":0,"adjustPoints":0,"foulPoints":0,"totalPoints":0},{"alliance":"Red","robot1Auto":"None","robot2Auto":"None","robot3Auto":"None","autoBouldersLow":0,"autoBouldersHigh":0,"teleopBouldersLow":0,"teleopBouldersHigh":0,"towerFaceA":"None","towerFaceB":"None","towerFaceC":"None","towerEndStrength":0,"teleopTowerCaptured":false,"teleopDefensesBreached":false,"position1crossings":0,"position2":"A_Portcullis","position2crossings":0,"position3":"C_Drawbridge","position3crossings":0,"position4":"D_RockWall","position4crossings":0,"position5":"B_Moat","position5crossings":0,"foulCount":0,"techFoulCount":0,"autoPoints":0,"autoReachPoints":0,"autoCrossingPoints":0,"autoBoulderPoints":0,"teleopPoints":0,"teleopCrossingPoints":0,"teleopBoulderPoints":0,"teleopChallengePoints":0,"teleopScalePoints":0,"breachPoints":0,"capturePoints":0,"adjustPoints":0,"foulPoints":0,"totalPoints":0}]},
Next year I plan to use TBA exclusively.
To give an idea of perspective, there were a total of 13302 matches this season, meaning 5.06% of all matches this year were decided by foul points (0.65% as Tiebreakers, 4.4% as Sways)
How does that compare with previous years?
How does that compare with previous years?
I don't have a database for previous years, unfortunately. Each year the schema changes due to a new game
How does that compare with previous years?
Without any data... not quite sure. I'd bet that 2014 had a bunch of Sways, though. Something about 50-point fouls...
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.