# Event box scores

I’m always interested in ways to present data in the most concise fashion that lets people extract information for their own use. I’ve followed the work of Edward Tufte, he has written some amazing books on the display of quantitative information.

Which brings me to today’s topic, robotic event box scores. We’ve all seen sports box scores that give the full information about a contest at a quick glance.

We have our own:

1. 225, 341, 304, 4637 W
2. 1676, 222, 2607, 224 QF
3. 1168, 272, 1712, 3929 QF
4. 1640, 1626, 2539, 708 QF
5. 2729, 11, 1403, 834 SF
6. 1218, 4575, 484, 4281 SF
7. 2016, 3314, 369, 1923 F
8. 365, 316, 869, 423 QF

(sample from Ramp Riot 2013, thanks Ben Martin!)

Because we know how the alliances are selected, we know how teams were selected. We also know who placed where to some extent. We don’t know the real details, unless we get this:

1 beat 8 in 2 matches
5 beat 4 in 2 matches
7 beat 2 in 2 matches
6 beat 3 in 2 matches

1 beat 5 in 3 matches
7 beat 6 in 2 matches

1 beat 7 in 2 matches

I’m looking for a way to add the match info to the short summary

Like:

1. 225, 341, 304, 4637 W (2,3,2)
2. 1676, 222, 2607, 224 QF
3. 1168, 272, 1712, 3929 QF
4. 1640, 1626, 2539, 708 QF
5. 2729, 11, 1403, 834 SF (2)
6. 1218, 4575, 484, 4281 SF (2)
7. 2016, 3314, 369, 1923 F (2,2)
8. 365, 316, 869, 423 QF

If you have 1 number at the end on your alliance, you won your quarter final in X matches. If you have a second number, that’s how many matches it took in the Semi’s. Three is the winner and how many matches their march to victory took.

Since the elimination tree is always the same, we know that 1v8, 2v7, etc for the quarter finals. So figuring their match ups are easy.

Is there a better way to represent this data? Is there missing data that would be helpful? For example would it be worth know in the top 8 seeds? And given just the data shown, could you reconstruct the eliminations?

I could.

#1 over #8, 2 matches
#5 over #4, 2 matches
#7 over #2, 2 matches
#6 over #3, 2 matches

#1 over #5, 3 matches
#7 over #6, 2 matches

#1 over #7, 2 matches

I would be interested in knowing original seeds; I think that could be done something like:

1. 225 [1], 341 [X], 304 [Y], 4637 [Z] W (2,3,2)

I use brackets instead of parentheses because the latter are used for the matches/round info, so I’d like to avoid confusion; X, Y, and Z are placeholders for the seed where it isn’t known (obviously the #1 AC is the #1 seed).

I agree, I think original seed is more valuable than how many matches it took. The other thing that’s not given and probably impossible to put in such a small format is match score. For all other sports that’s one of the most important items.

New proposal (fake data)

1. 225(1), 341(3), 304, 4637 W [2,3,2]
2. 1676(2), 222, 2607, 224 QF
3. 1168(4), 272, 1712, 3929 QF
4. 1640(5), 1626, 2539, 708 QF
5. 2729(6), 11(7), 1403, 834 SF [2]
6. 1218(8/5), 4575, 484, 4281 SF [2]
7. 2016(9), 3314, 369, 1923 F [2,2]
8. 365(10), 316, 869, 423 QF

Note 6, it shows that team 1218 was the 8th seed, and declined the 5th alliance. Trying to stay minimal, could be:

1. 1218(8D5), 4575, 484, 4281 SF [2]
2. 1218(8X5), 4575, 484, 4281 SF [2]

The other thing that’s not given and probably impossible to put in such a small format is match score. For all other sports that’s one of the most important items.

I’m spazzing out on sports that play mini-matches. I know tennis, best out of three, so the scores for the final match could be presented as:

1. 225(1), 341(3), 304, 4637 W [2,3,2] 72-65, 84-83
2. 1676(2), 222, 2607, 224 QF
3. 1168(4), 272, 1712, 3929 QF
4. 1640(5), 1626, 2539, 708 QF
5. 2729(6), 11(7), 1403, 834 SF [2]
6. 1218(8/5), 4575, 484, 4281 SF [2]
7. 2016(9), 3314, 369, 1923 F [2,2]
8. 365(10), 316, 869, 423 QF

Not sure that it would be worth to do all the eliminations, it’s starting to get long.

1. 2220, 2177(4), 3755
2. 2538, 3883, 4181
3. 2526, 2705, 3102 [78-68,100-67]
4. 4054, 2977(5), 2264 [58-18,59-50]
5. 4009, 4228, 4703
6. 4215, 1816(8), 93
7. 2052(d2), 2062, 4011 [115-71,154-52],[99-82,0-85,115-78],[94-105,122-76,126-90]
8. 3276, 3277, 4539 [99-45,123-61],[55-36,109-34]

The numbers next to the second team represent their rank at the time of their selection. The d2 means that 2052 declined the 2nd team. I felt that their position at the time does not need to be added here since, if you really want to know, it can be reconstructed from the given data. The match score sets show the rounds where that alliance advances. The QF, SF, F, and W are unnecessary since the match scores are given.

or maybe this:

1. 2220, 2177(4), 3755 [45-99,61-123]
2. 2538, 3883, 4181 [71-115,52-154]
3. 2526, 2705, 3102 SF [82-99,85-0,78-115]
4. 4054, 2977(6), 2264 SF [36-55,34-109]
5. 4009, 4228, 4703 [18-58,50-59]
6. 4215, 1816(10), 93 [68-78,67-100]
7. 2052(d2), 2062, 4011 W
8. 3276, 3277, 4539 F [105-94,76-122,90-126]

This one looks much cleaner but is much less intuitive. The score sets come from the round of the tournament where the alliance was eliminated.

One of the problems I see is that you are not always going to have all the information. Many times these are put together by people that just happened to be at the event and collected from data after the fact. Allowing for a incomplete data set is useful. In this example someone knew a lot about the rankings and was able to fill them in for many of the teams but only watched on side of the elimination bracket.

1. 225(1), 341(3), 304(25), 4637 [W x,x,2] {F 72-65, 84-83}
2. 1676(2), 222(15), 2607(31), 224
3. 1168(4), 272(12), 1712, 3929
4. 1640(5), 1626(18), 2539, 708
5. 2729(6), 11(7), 1403, 834 [S x]
6. 1218(8/5), 4575(11), 484, 4281 [S 3] {QF 64-30, 46-62, 43-25}
7. 2016(9), 3314(13), 369, 1923[F 2,2] {QF 78-30, 80-32} {SF 32-30, 81-42}
8. 365(10), 316(14), 869, 423