Statistical Analysis of Regional Competion Scores

[rourke]

I compiled some basic statistics on the qualifying match scores of Regional Competitions through March 20th from the FIRST Regional Events Match Results web pages. Only qualifying match scores were used. Elimination round matches were not included.

The average score of ALL qualifying matches is 51 (you “hangers” have an above average chance of winning if you can do it consistently!). The standard deviation is 34. The highest qualifying score has been 200. Only New England has not had a “zero” score (lowest score is 5). 83 percent of all scores are less than 85. 97.5% of all scores are less than 120. New England’s average score of 72 is significantly higher than the other Regionals. Pittsburgh has the lowest average score of 41. The Arizona match results aren't posted yet, so I don't have any stats.

How does your robot stack up against these stats?

NH NJ VCU OR AZ
Avg. Score 56 51 43 49 N/A
Std. Dev. 37 32 29 35 N/A
High Score 150 140 125 170 N/A
Low Score 0 0 0 0 N/A
83% Score 93 83 72 85 N/A
97% Score 131 115 102 120 N/A

Regional: FL Gr. Lak. Pitt St. L NewEng.
Avg. Score 52 50 41 45 72
Std. Dev. 33 32 29 30 42
High Score 140 165 120 145 190
Low Score 0 0 0 0 5
83% Score 86 81 70 75 114
97% Score 119 113 99 104 156

Regional: Ches. Detroit Peach Sacr. Long Is.
Avg. Score 54 52 43 48 57
Std. Dev. 34 32 30 35 34
High Score 185 140 160 200 155
Low Score 0 0 0 0 0
83% Score 89 84 74 83 91
97% Score 123 116 104 117 125

Steve Rourke
Mentor / GM Engineer
Team #1114 GM SIMBOTICS

[KenWittlief]

interesting! one of our mentors, Dr Hensel from RIT, did a statistical analysis of the potental scores right after the kickoff

and concluded that getting above 100 points would almost guarentee a win - in other words, most teams would score less than 100 points

which has proven to be correct.

Math is our friend.

BTW - based on his analysis we designed our bot to be able to get >100 points per match, even without a functioning alliance partner.

[Jay H 237]

One of the things I noticed at UTC is that the red alliances won more than the blue alliances. I don't know the exact percentage red had over blue but it seemed like if you were picked to be blue you had less of a chance at winning. There seemed to be a pattern after awhile.

[KenWittlief]

do you mean in the elimination rounds? teams stay on the same side of the field during the elim rounds, so whichever alliance wins, then yes - that color team did win more matches than the other.

logically one side of the field will win more matches in the seeding rounds too - it would be strange if 80 matches were played and blue won exactly 40 times.

[Karthik]

It will be interesting to see how the numbers pan out in weeks four and five. For the most part, teams who were competing in weeks 1-3, were out there for the first time. Traditionally weeks 4-5 are where scoring picks up. From there scores always zoom upwards at nationals (Take a look at 2001 for a good example).

As for the New England anomaly, it would be easy just to attribute it to the fact that 50% (19/38) of the teams have at least six years of experience. But if we dig deeper, we notice that small balls were more of a factor. If you look back to 2002, we recall that many of the best ball handlers came from the New England area. I think this is why the scores were dramatically higher at the UTC event, teams were experienced with the small balls.

If you look at what the Cheesy Poofs did in Sacramento (http://www.chiefdelphi.com/forums/sh...threadid=26761), you see that their success can be attributed to the fact that they had time to gather balls, cap the stationary and go hang. Those extra small balls took them from being a 100pt team, to being a 150pt team.

As the season progresses, look for more teams to try and more than just cap and hang, and as this happens look for scores to go up.

[Ericente]

Yeah, I was at UTC New England, and it was a excellent, very high scoring competition. I attribute it to the fact that there were a lot of vereran teams and that most of the teams had already been to another regional and have had time to practice.

BTW, I believe New England's highest score was 200 points by the Gael Force alliance in the semifinals, not 190.

[Solace]

ok, so i actually did crunch our numbers here

571 Paragon
13-2-0 at UTC

average score, 112 pts
opponents average score, 75 pts
average winning margin, 49 pts

average score in the finals, 150
median score in finals, 160

my number crunching abilities are far from perfect, but that's about it

[MisterX]

Originally Posted by Ericente

Yeah, I was at UTC New England, and it was a excellent, very high scoring competition. I attribute it to the fact that there were a lot of vereran teams and that most of the teams had already been to another regional and have had time to practice.

BTW, I believe New England's highest score was 200 points by the Gael Force alliance in the semifinals, not 190.

I believe Rourke said he calculated using only qualifying and not the actual elimination so the semifianls match was not tabulated in any of the average highs, means modes or any of the such. But yeah that was the highest of the regional just not of the tabulated ones

Team 521 Stats Through qualifying matches
6-3 QP 61.1
Average score 82.8
Average winning score 101.7
Average winning margin 43.3
Highest score: 135
Average # of balls in goal: Plenty especially with the multiplyer

[Joe Ross]

After you've recovered from the canadian regional, can you update this thread?

[Karthik]

Originally Posted by Joe Ross

After you've recovered from the canadian regional, can you update this thread?

Joe,

I don't think I'll ever recover from the Canadian Regional. Nevertheless, here are the updated regional stats...

Event: OH CO IL NY PA CA
Avg. Score 54 44 65 44 70 47
Std. Dev. 36 31 40 32 42 35
High Score 175 180 180 145 190 170
Low Score 0 0 0 0 5 0
83% Score 90 75 105 76 112 82
97% Score 126 106 145 108 160 117

Event: ON TX SC SJ MI
Avg. Score 45 49 55 51 73
Std. Dev. 32 30 33 39 42
High Score 150 130 130 155 180
Low Score 0 5 0 0 0
83% Score 77 79 88 91 115
97% Score 109 100 121 129 157

If some of the 83% and 97% scores look off, it's probably because of rounding. I'm working on getting a full spreadsheet of regional data posted soon. I may even sort everything by division as well.