Unofficial Team Tracker

2 replies [Last post]
dgoree
Title: NooBot
Joined: 02/13/2012
Posts:
BotPoints: 130
User offline. Last seen 1 year 12 weeks ago.

This year's Official Team Tracker has done a great job posting and analyzing results from tournaments where there are enough volunteers to receive and process official scores, but that still leaves most regions unreported. This year I am working on a spreadsheet to back-work the official over-all results into the maximum seeding scores and seeding averages at each tournament. Unfortunately, I am running into a few problems. Back-working yields several possible sets of scores and I can't always confirm which set is legitimate. Of the regions that have taken place so far, I can't confirm Texas or Arizona scores. Is there anyone that knows any seeding average from those regions that isn't 0? If I get just one non 0 score confirmed I can get the rest.

If you are interested in looking at what I have so far, click this link: https://docs.google.com/spreadsheet/ccc?key=0Aq9uurSNlyXZdDFwVXpwYks4X21...

I have allowed public editing for almost everything if there is a score you would like to confirm or something you want to add. I do ask, however, that you don't change any of the formulas or locations of data. The play sheet is there as a sand box if you want to play around with any of the numbers. If you know a seeding score for one of those unconfirmed regions, increment or decrement the H6 box by 0.5 until you see the correct scores, and then please put a comment there so everyone knows they are accurate. Comment anywhere and everywhere you please.

-Daniel Goree
Former...
Senior Project Manager
President
Norman Advanced Robotics (13-0113)

Marty Rand
Marty Rand's picture
Title: Botball Youth Advisory Council
Joined: 07/04/2009
Posts:
BotPoints: 253
User offline. Last seen 3 years 23 weeks ago.

One of the things we plan on doing is getting official scores from KIPR. I didn't mention it but we have a way of inserting the released score data from (almost) every region. This would give us overall rankings for those regions which we cannot cover live. We already have Arizona's and Texas's data in there. We feel that while you can extrapolate the scores for a team, it is better to get official data from KIPR. KIPR has been reportedly saving all of the score sheets from every region for us. This will let us get official Seeding, DE, and Alliance scores while your method cannot get the 3rd seeding score for a team or the actual DE match scores. Our method can also make a bracket for DE (a feature we are planning). And because of the nature of how you are going it, there will probably be some errors since there are multiple ways to get a seeding component of 0.9. I am very curious as to how your scores and our scores match up. Also, I noticed your spreadsheet doesn't have every score in it, do much statistical analysis of the data, have DE / Alliance matches in it, or update in real-time for some regions (just to name a few). There is a reason the Team Tracker page says "EARLY BETA" on it. It is still evolving and maturing. Plus, I have to code everything, not plug in data and watch Google do the rest.

-Marty Rand
{
Senior programmer at Norman Advanced Robotics

Former senior programmer at Whittier Middle School

Youth Advisory Council

All around nerd
}

dgoree
Title: NooBot
Joined: 02/13/2012
Posts:
BotPoints: 130
User offline. Last seen 1 year 12 weeks ago.

Oh no the Official Team Tracker is definitely a better system than what I have. I agree that official scores from KIPR are the way to go, and yes, those spreadsheets provide a limited amount of information. I am working on what I have as a temporary project to provide seeding scores while the Official Tracker is still in "EARLY BETA" stage. By no means will what I have be worth anything once the Official Tracker is completed.

Also, this was originally intended for personal use for the specific things I was interested in, and as of yet I have not found need for the other features. I only posted it here because I felt that it would not be fair to ask for information to help validate my data if I did not share the data with those who might help. This is in no way shape or form designed to compete with the Official Team Tracker. It is just to provide scores that the Official Tracker does not, at this time, have access to.

Here is basically how my method works:
The overall seeding score formula has two unknown values at the time of the official score postings: each team's average score, and the tournament maximum score. Due to the nature of this year's scoring and the mechanics of a two item average, every seeding score must be divisible by 1/2 of the lowest possible score. Last year that was 2.5, this year it is 0.5. To find the tournament maximum score, I assume that the second lowest seeding average is that number (0 is pretty much always the lowest score and can't be solved for). With the maximum score found, I can use basic algebra to find each team's seeding average. Naturally, that 0.5 score is not always the second least score. In order to check that, I look at the computed averages. If the scores are not divisible by 0.5, I know that 0.5 is not the second lowest score, and I increment by 0.5 to 1, then 1.5 ect until I have a set of potentially valid seeding averages. The reason that these are only potentially valid scores is because any second lowest score that is divisible by that first score will produce potentially valid results. At this point I look critically at the scores to see where the scores make the most sense. For example in Oklahoma, a value of 0.5 makes the max score 880. A value of 1 is 1760. That is unreasonable and not valid. Therefor, I can conclude that my set of Oklahoma scores is reasonably reliable.

My errors in this data will only be present in that second lowest score. If that is right, everything will be correct. If it is incorrect, the specific scores will be wrong, however the 0 scores are correct (they do not depend on the maximum score) and the ratio between the non 0 scores is still correct. To check that, go to the Texas sheet, and change the H4 value to any multiple of 2.5. You will notice that the shape of the graph of scores remains unchanged. So if you were to compare my data to the Official Tracker's, you would see the same averages as I would have selected my H4 value to match your set of data, or some other confirmed score report (like the Florida results pdf)

The reason I started this thread was not to try to compete with the Official Team Tracker, or increase traffic to my project. I just wanted to see if someone could help me confirm a portion of my project. I am using it for personal experimentation in spreadsheet design, region comparison, and graphing of scores. (Interestingly enough, as initially noticed by I believe Kyle Montgomery at GCER 2011, scores are generally following an exponential trend. If that continues to hold, we can expect scores in the 200s to 600s to show up in the next few regions.)

I am thrilled at the improvements coming to the Official Tracker, and am particularly excited at the potential of comparing scores from year to year as well as following distant tournaments as they happen. What I have now isn't even close to as cool as what the Official Tracker will have to offer. It is just a personal project to find scores that are not yet available as the Official Tracker continues to grow.

-Daniel Goree
Former...
Senior Project Manager
President
Norman Advanced Robotics (13-0113)