Currie SCAM Ranking

[rourke]

I’d previously published a list of all team qualifying average scores greater than 90. There were 50 teams on that list. But since there was a large variation in Regional Event averages (eg. 41 @ Pittsburgh; 73 @ West Michigan), comparing robots solely on the basis of average score is misleading…..So, I tried to neutralize the effect of Regional differences with an equation I called “Scoring Capability Analysis Model”. The acronym SCAM was purposely chosen so that no one would take this too seriously.

The SCAM number was calculated using the ratio of Team Average (TA) to Regional Average (RA) and makes an adjustment for performance consistency using the Team’s Standard Deviation (SD). The equation identifies teams that had high scoring averages relative to the other robots they competed against, and who also demonstrated consistently high scores – without a large variation from match to match. If you find yourself paired with one of these robots, expect a high score. If you’re competing against them, you might want to consider defense.

SCAM = (TA-SD)/RA

The Curie Division Top 10 SCAM Robots are:

1. 522
2. 233
3. 1114
4. 279
5. 66
6. 57
7. 11
8. 977
9. 292
10. 61

Honorable Mention goes to Teams 74, 224, 1243, 126, 175, and 302 (who all had scoring averages over 90, but didn’t make the top 10 SCAM list).

[RoteAugen]

out of curiosity.. which teams have the highest QP avgs.? by that i mean the win/tie/loss point averages (2=win, 1=tie, 0=loss)

It's not that I think the regular point averages are unimportant, but it is my experience that the teams that win more often do not necessarily score that many points, especially if defense is a part of their strategy

[Joe Ross]

Originally Posted by RoteAugen

out of curiosity.. which teams have the highest QP avgs.? by that i mean the win/tie/loss point averages (2=win, 1=tie, 0=loss)

JVN posted a white paper which shows this.

Rourke, Can you post the actual value of the SCAM for the top teams? Did anyone get a value above 2?

[rourke]

Rourke, Can you post the actual value of the SCAM for the top teams? Did anyone get a value above 2?[/quote]

1. 522..........1.65 pts
2. 233..........1.59
3. 1114.........1.55
4. 279..........1.33
5. 66............1.2
6. 57............1.14
7. 11............1.13
8. 977...........1.1
9. 292...........1.07
10. 61............1.0

Note - Curie has the highest SCAM teams of all the divisions

Galileo's highest is 1.34 (Team 5)
Archimedes' highest is 1.4 (Team 945)
Newton's highest is 1.6 (Team 254) (and then the next closest is 1.2)

[mtaman02]

Curie has got the highest SCAM teams ?

If you don't mind me asking what do u mean by SCAM and is it a good or bad thing.

otherwise nice stats for the top 2 robots in that last post (522, 233)

I don't see how 254 isn't in the top 5 though. they have had many successful yrs as well as this yr all by itself.

oh well we shall see in but just a few days on who will win what which reminds me

In the first curie thread. I would like to start up like a "ventures related thing)

CD members may post who will win what award and the chances of them winning that said award. Same goes for the semi / quarter / finalist in both division and main national event

no fancy program to do all the figures just a normal post

[Aaron]

I was going over the SCAM system and i have to say it is probally the best mathimatical way to rank teams. Out of curiosity did u use every regional becaus i know that we did really bad at Trenton but really well in both S.C. and Annapolis.

[rourke]

Originally Posted by Aaron

I was going over the SCAM system and i have to say it is probally the best mathimatical way to rank teams. Out of curiosity did u use every regional becaus i know that we did really bad at Trenton but really well in both S.C. and Annapolis.

Like most teams that competed in multiple events, Team 11 improved from event to event. All teams and all regional events were evaluated - and your best was in SC (your third event). There were dramatic improvements in your average score (44 in NJ, 66 in MD, and 91 in SC). Your scoring variation (standard deviation) also reduced by 13 points over the 3 regionals - which means you are scoring more consistently too. The combination of high scores and scoring consistency is why you landed in the top 10. You will be a formidable competitor in Atlanta if the trend continues! Looking forward to competing with you. (Hopefully as an alliance partner, of course)

[Karthik]

For anyone who is looking for SCAM ratings for all teams this season, take a look at the white paper I uploaded yesterday. http://www.chiefdelphi.com/forums/sh...ad.php?t=27656The SCAM ratings are available in the column labeled "factor".

[Driesen]

Originally Posted by rourke

Like most teams that competed in multiple events, Team 11 improved from event to event. All teams and all regional events were evaluated - and your best was in SC (your third event). There were dramatic improvements in your average score (44 in NJ, 66 in MD, and 91 in SC). Your scoring variation (standard deviation) also reduced by 13 points over the 3 regionals - which means you are scoring more consistently too. The combination of high scores and scoring consistency is why you landed in the top 10. You will be a formidable competitor in Atlanta if the trend continues! Looking forward to competing with you. (Hopefully as an alliance partner, of course)

Wow you really know what your talking about here. I knew we improved a lot, but by your numbers I didn't realize how much. There's a lot of good robots in our division there definetly will be lots of good competiton. I have heard nothing but great things about your team, and definetly look forward to compete with you guys as well. See you in Atlanta!

[Stephanie]

Originally Posted by mtaman02

If you don't mind me asking what do u mean by SCAM and is it a good or bad thing.

read the first post in this thread

Originally Posted by mtaman02

I don't see how 254 isn't in the top 5 though. they have had many successful yrs as well as this yr all by itself.

254 isn't in the top 5 because there are other teams that performed better than them this year.

[Cory]

Read the first post in this thread. a high SCAM is a good thing.

Team 254 isn't on that list because theyre in Newton, not Curie.

Cory

[rourke]

Originally Posted by Cory

Read the first post in this thread. a high SCAM is a good thing.

Team 254 isn't on that list because theyre in Newton, not Curie.

Cory

If you are interested in seeing the top SCAM robots in each division - there are separate threads for each. Team 254 is in Newton - and they're the top SCAM robot in Newton.

The top 30 SCAM robots from all divisions are (in order): 522, 233, 254, 1114, 945, 5, 279, 56, 1305, 494, 222, 111, 311, 60, 814, 237, 1030, 66, 569, 48, 486, 456, 33, 57, 67, 11, 340, 1403, 141, 281.

[rourke]

Originally Posted by RoteAugen

out of curiosity.. which teams have the highest QP avgs.? by that i mean the win/tie/loss point averages (2=win, 1=tie, 0=loss)

It's not that I think the regular point averages are unimportant, but it is my experience that the teams that win more often do not necessarily score that many points, especially if defense is a part of their strategy

Maybe, but consider this.....

Of the SCAM robots listed above, between them there are 13 Regional Winners or Finalists! (some multiple times) These 'bots seem to know how to score AND win too.

[RoteAugen]

Team Number........Rank .......Average Seeding Points
340.....................1 ........2
501................... 2........ 2
967................... 3........ 2
1405...................4........ 2
233................... 5........ 1.85
179................... 6........ 1.8
60..................... 7........ 1.778
222................... 8........ 1.778
360................... 9........ 1.778
492................... 10........ 1.778
522................... 11........ 1.778
945................... 12........ 1.778
1359................. 13........ 1.778
254................... 14........ 1.764
138................... 15........ 1.75
311................... 16........ 1.75
538................... 17........ 1.75
27..................... 18......... 1.714
292................... 19........ 1.714
859................... 20........ 1.714

those are the stats i was looking for, if anyone else wanted to see them

[Zzyzx]

Originally Posted by RoteAugen

492................... 10........ 1.778

Hate to break it to you, but we are in the Gallileo division, not Curie.