ONLY AT THE MICHIGAN STATE CHAMPIONSHIP

FIRST General Forum
#26

Signal to Noise Ratio is used a lot in engineering methods were physical phenomena are measured for quality. It is used in communications, electonics and recently is used a lot in quality methods to measure the quality of processes and components. SNR (or S/N) is a logarithmic value measured in dB.

It turns out that methods for measuring quality in the world of Engineering and Manufacturing can be easily applied to sports. Another great opportunity for crossover learning in the world of STEM/FRC.

SNR give you a ratio of variabliy compared to the level of the signal you are measuring. In this case I am using Average Event OPR as the measurment of interest. The higher the SNR value, the more “competatively balanced” the event is, meaning that the distribution of the capability of teams is tighter than at an event with a lower SNR value.

Example, Compare Central Valley with Alamo. Both have nearly identical average OPR numbers for the event, (15.4, 15.2), however the distribution of OPRs for the teams at Alamo was much broader than at Central Valley.
What this a higher SNR means is that the event not likely as dominated by a few good teams, but instead was much more balanced competitively.

On these scatter graphs, Higher is means better overall event scoring ability, points to the right have better competitive balance.

I plan to write a paper on this, if I can ever find the time.
I use this same method to measure many factors in FRC competitive analysis, from the quality of individual teams, team quality growth over time, game design comparisons, and more. Fun with math.

#27

The point of IRI is to see the most instense competition of the year, thats why you have to be crazy good to get in…

#28

So what is it, the standard deviation of OPR in dB or something similar?

#29

S/N = 10 * Log ( Average^2 / StdDev^2 )
This is a standard formula from Signal Processing and related fields.

Events with tighter grouping will have a higher SNR.
Images below depict this graphically.


#30

The great thing about IRI is that you lack the exclusivity of MAR or MSC - that is, you have the best MAR and Michigan teams plus the rest of the best from the rest of the world. It’s absolutely no contest that IRI is far more competitive than either of the regional championships. With that said, comparison between Divisional competition and Einstein vs. MAR or MSC eliminations can certainly fluctuate. Take, for example, the eliminations in 2011. I can’t speak to the rest of the divisions, but eliminations on Galileo were fairly intense, but the #2 alliance of 254, 111, and 973 won handily, and were hardly challenged on Einstein. Except for one match in the Galileo finals, there were no upsets and very few close matches. In that year, at least, I would consider MSC’s elims to be far more exciting (Whenever a #8 seed alliance can upset a #1 alliance like 217, 469, and 201 can be upset, you know the event must be deep). That said, neither events held a candle to the IRI.

#31

The part I’m surprised by is that MSC is SO much more competitive than MAR.

#32

IRI is insane. You need to be there to see it.

#33

I wonder if this has anything to do with how many MAR teams turned down their MAR Championship invitation, as they qualified at a regional? I don’t think MSC had any teams turn down that invitation.

Quick note:
A team from the Upper Peninsula (at least a 10 hour drive) found out on Tuesday that they qualified for MSC. They were able to pay and organize in time to be there Thursday morning. They passed inspection quickly. I haven’t had time to look up how they did, but I was so excited that they made the effort to come down for the event. What an awesome experience for their team!

#34

IRI, no doubt.

#35

Unfortunately Team 216 had to turn down the invitation to states. Our robots were still making their way through customs after competing as a finalist at the Western Canada Regional in Alberta. If there was any way for us to have attended, we would have!! So sad that we had to miss this crazy Championship! :frowning:

#36

857 ranked 56th, with a 4-8-0 record. It’s about a 9 hour drive to MSC for them, and they probably encountered at least a little snow somewhere along the way (either on the way there or on the return trip).

#37

I remember when we competed at Waterloo this year and seeing the high scores, worlds highest average winning score, and how competitive almost every team was. Watching MSC blew that out of the water. MSC was just insane. I don’t see Champs Quals being higher but elims will be probably 150-175 Average Winning Score.
IRI is IRI. That is all.

#38

It doesn’t seem that MSC has a necessarily tighter grouping, just that it’s mean shifted up by about 20. That’s something I’d expect when you’ve created a performance filter on teams.

Compare that to the Championship event, where each regional produces 6 invitations; where 2 of the invitations are filled by the 2 most competitive teams and the other 4 aren’t necessarily; unless the event was deep and the 20 something pick was still a good robot.

#39

Also. STOP CALLING IT NATIONALS! Doesn’t World Championship have a stronger ring to it anyway?

Its an international competition, and it hasn’t been called nationals since 2002, when the Canadian Regional (now Greater Toronto West) started 11 years ago. In the time since, there is now 6 regionals (5 in Canada and 1 in Israel) that happen on international soil, and others in the works (AFAIK Australia and Mexico are on the fast track). 234 International teams (of 2538 registered teams in 2013 [9.2%]) representing 15 countries compete in FRC. To call FIRST Championship “Nationals” makes us feel like we don’t exist.

#40

One of my pet peeves too, although I generally just cringe and deal with it.

#41

Don’t.

You can’t expect anything to change if you’re not going to step up and try to make that change. Even if that means calling out everyone you see/hear making that mistake. As Racer26 pointed out, it’s been a WORLD Championship for over a decade. As a community, we have no excuse for continuing to call it nationals.

#42

Mr. Zondag,

Thanks for the information on the signal noise ratio, I will be showing my students tomorrow. It reminds me of Cp and Cpk in determining process capability, but with those you have upper and lower spec limits to base off of. Here we don’t so I’m glad to know a way to “judge” the equality of the competitors.

#43

Isn’t higher score means poor defense? If you watch some of the matches you see robots hurrying like mouse scurrying with a cheese bit!

Hats off to the drivers! Do the experience carpal tunnel syndrome?::rtm::

#44

Check out the elimination matches. While the quals are usually an offensive burst by all teams, eliminations is where defense usually shines. I don’t think anyone normally tries to play defense during the quals unless you’re in MAR, a defense heavy region.

#45

This kind of reminds me of the Sharp ratio for portfolio selection. So FIRST is relevant to engineering, sports, and now economics!

Also there were a total of four FiM teams that turned down MSC; 216, 1596, 94, and 815. Don’t know anything about MAR. 216 has already stated that they were stuck in international shipping. I’d be interesting to see if any of the other 3 teams had nonfinancial reasons for not attending.