Practice bot morality

[AdamHeard]

Originally Posted by nickdog8891

I couldn't disagree with you more. Whole-heartedly, unequivocally disagree. And not just in regards to FIRST

My point was that one can always improve their circumstances through hard work (applied in the right direction, hard work the wrong way does no good).

Unfair is only brought by people who are unwilling to do the above.

I'm not claiming everyone can become a billionaire, astronaut or other wildly lofty goal merely through hard work. I'm claiming that one can always improve through proper application, and therefore there is no unfair.

My team is a perfect example. We used to be a have-not, and now we're a have. This was achieved exclusively through hard work.

[IKE]

Once upon a time, there was a competition called the Fair Robotics Compeition. The principle of the competition was fairness first (even before safety). It was the fairest compeition in all the land. The one guiding rule was that fairness was in the eye of the beholder, and thus if someone thought something was unfair, the Fair Robotics Compeition would make a change. The first year of the competion was a lot like the FIRST Robotics Competition, but there were a ton of complaints. The next year, practice bots, and going to more than one compeition were dissallowed. Everyone noted that the scoring was much lower, and the action much worse, but it was more Fair, but still not fair enough. Many teams were larger and had more money, so the Fair Robotics Compeition board put a cap on total team budgets. There was also a lot of complaints about tool useage, so there were strick enforcements of which tools would be allowed. The actual compeition at the event suffered more. Most agreed that it was more fair, but there were still a few naysayers that thought it was unfair that some teams had more man-hours because they were larger. Also, not every team had a technical mentor, so those were not allowed. Lastly, not every team had a programmer, so only base code was allowed. That year, the robots really suffered, and scoring got even lower. By this point, the GDC had lowered the bar to the most rudimentary tasks, but without any technical mentors, most of the robots suffered to drive around much at all. The following year, Texas got hit with a Blizzard, that made everyone have to stay home. In order to be fair, a temporary stop work was placed on all teams in order to be fair. Unfortunately, it was the last week of build season, so virtual no robots were ready for compeition. That year really sucked. Because of it, the GDC got together and decided the next year, teams would build and assemble their robots at the compeition. In order to do this, there robots were greatly simplified. The task was also greatly simplified to essentially driving around a course on the carpet. While everyone agreed the competition was extrememly fair, it was universally agreed that it was incredibly un-inspiring. Almost everyone left the Fair Robotics Compeition for something more interesting and exciting. It was called the FIRST Robotics Compeition. While not as fair as Fair Robotics Competition, it was a lot more challenging, exciting and inspiring.

In racing, there are tons of series that try to promote "fair" racing. The tighter the control, then the closer the field is. The closer the field is, the more powerful tiny "cheats" become. I raced in Spec Neon for a few years. At the event I raced at, a good time was around 1:21 to 1:22s per lap. A 1.25% cheat would remove basically 1 second from your lap time. For a car that produces 138 HP, this was finding an engine cheat that would increase performance by just 1.7 HP. That 1 second was usually the difference between 1st place and around 5th place. SCCA racing is even worse. If you do too good with a particular car, it may get promoted to the next class where it suddenly is a slow worthless piece of junk.

I would recommend spending less time worrying about "fairness" and more time worrying about your team achieving its objectives. If your goals are wanting to perform at a really high level with respect to the peers on the playing field, than you better learn what they do that makes them perform well, and compare/contrast that to your program. If your objectives are to learn some neat stuff and not really concerned about the outcome, great. One thing I do not understand though, if a team's goal is just to show up and learn a bunch of neat stuff, why would they care if a compeition was "fair"? Wouldn't they just be concerned with whether or not it was a good learning platform?

Back slightly on topic, trying to produce a second robot "practice" robot that acts the same as your competition robot is very difficult. There is a ton of learning that comes out of just trying to get the two to act the same...

[Justin Montois]

Well said IKE.

/Thread perhaps.

[SteveGPage]

What he said! ^^^^

[BrendanB]

Amen IKE!!!!

I can bet that everyone who says practice bots are unfair doesn't have the resources to build one. If they pushed their team harder to get those resources, completely different story.

[Ninja_Bait]

Just to add one more insight along the lines of IKE's, from our friend Kurt Vonnegut: http://www.tnellen.com/cybereng/harrison.html

[2544HCRC]

I've also been involved in car racing. I take a slightly different view. What happens in most racing series is that a few teams start to dominate the field because of resources. Pretty soon all of the other teams start to think "what's the use?" and quit showing up and pouring money down an unfillable hole. So racing groups including SCCA institute a class system. Imagine racing without the class system. You would have far fewer racers. It turns out it's much more fun when you show up and have a shot at winning. If the rules aren't adjusted the field dwindles and dies. I see that in FIRST. Rookie teams are sold one thing and show up to another. My guess is that if the presentation for FIRST went something like you are going to need a team of Engineering mentors, access to a pretty sophisticated shop, and the finances to build 2 robots + in order to be competitive, the field would and will be much much smaller. We did BEST this year as a trial and will probably switch over to FTC and BEST next year. For us as a small team with kids that aren't interested in building a big team and very limited resources, FRC has proven to be just too much. If you have a successful FRC program, great but this entire discussion seems to hit a nerve that most of us feel and that is, robotics is great, the idea is awesome as a tool for teaching kids, but FRC might not be the best fit for many of the teams. It's only fun being a back marker for a little while. Pretty soon the newness wears off and you have to make a decision, do you want to do what it takes to win in the class your in or do you want to find another class that might be a better fit.

[Al Skierkiewicz]

Originally Posted by 2544HCRC

My guess is that if the presentation for FIRST went something like you are going to need a team of Engineering mentors, access to a pretty sophisticated shop, and the finances to build 2 robots + in order to be competitive, the field would and will be much much smaller.

I would like to point out that there is rather large number of rookies in picking position at each regional and many are on winning alliances. I would also like to remind folks of 2041 (coming from an inner city, largely Hispanic population) taking Silver medals with one robot, eight students total and only three engineering mentors in both the Minnesota Regional and 2010 Champs. And let us not forget my all time favorite hardworking team, 842, Falcon Robotics. When I think of what a team can accomplish, I think of what that team, student and mentors, are able to accomplish.

[XaulZan11]

Originally Posted by Al Skierkiewicz

I would also like to remind folks of 2041 (coming from an inner city, largely Hispanic population) taking Silver medals with one robot, eight students total and only three engineering mentors in both the Minnesota Regional and 2010 Champs.

I agree with your overal point, Al, but I think there are a ton of teams that would kill for one let alone three engineering mentors (especially one with the knowledge and experience that Rich Olivera has).

[lemiant]

Originally Posted by Al Skierkiewicz

I would like to point out that there is rather large number of rookies in picking position at each regional and many are on winning alliances. I would also like to remind folks of 2041 (coming from an inner city, largely Hispanic population) taking Silver medals with one robot, eight students total and only three engineering mentors in both the Minnesota Regional and 2010 Champs. And let us not forget my all time favorite hardworking team, 842, Falcon Robotics. When I think of what a team can accomplish, I think of what that team, student and mentors, are able to accomplish.

Al, this is one thing I think you are mis-representing. Yes the rookie class has elite teams and the veteran class has elite teams, but these two are not comparable. What is the highest team number ever to win champs? 1114.
Some awesome things can be done with a lot of dedication and very little resources but I don't remember the last time one of those brand-new, little teams knocked out your alliance at champs.

FIRST is H-A-R-D!!! And to compete with the best your going to need comparable resources.

[EricH]

Originally Posted by lemiant

Al, this is one thing I think you are mis-representing. Yes the rookie class has elite teams and the veteran class has elite teams, but these two are not comparable. What is the highest team number ever to win champs? 1114.
Some awesome things can be done with a lot of dedication and very little resources but I don't remember the last time one of those brand-new, little teams knocked out your alliance at champs.

FIRST is H-A-R-D!!! And to compete with the best your going to need comparable resources.

Two rookies made Einstein in 2010. In 2009, a single rookie made Einstein. Two years before that, in 2007, a sophomore team made Einstein (and that same team almost went the year before; their robot started having parts fail in division finals). And it took 2056 multiple years to make Einstein; they still haven't won it, despite having not lost a single regional they've attended since starting up. (They were Galileo semifinalists their rookie year.) If you're looking for winning the Championships as your measure of winning, 3/2353 teams will win it this year; that's 0.127%, which means it's pretty tough on just about everyone to win it. (Odds improve to about 0.88% if you just factor in the roughly 340 Championship teams.) Yes, those are under 1% chance for any given team.

[pfreivald]

Since when are any of us owed victory?

A couple of rambling thoughts on this meandering topic:

One of the coolest moments in 1551 history was when the Thunder Chickens were worried we weren't going to pick them at FLR in 2010. (We did pick them, and we won, and it was another of the coolest moments in 1551 history!) No one gave us that; we earned it, and it's more special because of it. ...and yet the trip to championship was near disaster on the field, thankyouverymuch, as we learned some important lessons about durability when you do multiple events. (Murphy camped in our pit in Atlanta, but everything that went wrong was, ultimately, our fault.)

There's nothing wrong with the bar being set very high (as in, as high as other teams choose to set it within the bounds of the rules), but there could be something to better educate the rookie teams on what they're getting into. Pulling in kids and adults from other districts for a year or two before spinning them off into their own team gives them a much firmer foundation, and a much better idea of what they're getting into.

Talk to the mentors of 217 and 254 and 1114 and 2056 -- they'll happily tell you what they've done to get to where they are. Use that information as you see fit, whether it's a team overhaul or incremental improvement. Don't bemoan the circumstances that put your school or team at a disadvantage, or do, but either way take them as a challenge and circumvent them as best you can.

[lemiant]

Originally Posted by EricH

Two rookies made Einstein in 2010. In 2009, a single rookie made Einstein. Two years before that, in 2007, a sophomore team made Einstein (and that same team almost went the year before; their robot started having parts fail in division finals). And it took 2056 multiple years to make Einstein; they still haven't won it, despite having not lost a single regional they've attended since starting up. (They were Galileo semifinalists their rookie year.) If you're looking for winning the Championships as your measure of winning, 3/2353 teams will win it this year; that's 0.127%, which means it's pretty tough on just about everyone to win it. (Odds improve to about 0.88% if you just factor in the roughly 340 Championship teams.) Yes, those are under 1% chance for any given team.

This is what I said. Rookies can be great, and I don't doubt that it is only a matter of time until one of them actually wins champs, but the odds ARE against them. Resources help you win, I don't think anyone here is going to challenge that.

And just for fun here's some math (Admittedly filled with all sorts of assumptions):

Last year, with numbers almost all the way to 4000, the un-weighted odds of a single team on the winning alliance being below 1000 was 25%... All three teams on the winning alliance were below 1000! The odds of that were 25%^3 or ~1.5%. that means that statistically the odds of all three teams on the wining alliance being below 1000 were just higher than the odds of any individual championship team winning. Obviously that is absurd! I could predict that the same will happen this year with at least a 50% chance of being right, you couldn't pick a winning team with anything approaching certainty.

Or take it one step further. 1114 was founded in 2003 so no team founded in a year after that would has ever won the championship. Assuming the ~2800 teams founded between 2003 and 2011 were founded in even increments of 350 teams per year the odds of none of these teams winning championships between 2004 and 2011 are:
((1200/1550) * (1200/1900) * (1200/2250) * (1200/2600) * (1200/2950) * (1200/3300) * (1200/3650) * (1200/4000)) ^3
or .00000054%.

I believe we can effectively determine from that number that veterans winning is not just statistical variation

[IKE]

While i generally applaud folks that do the math, with 2400 teams and numbers well into 4,300+, there is nearly a 50% total attrition rate in FRC. The probability of a team winning the 2010 championship when they were a one year wonder in 2008 is 0%...
********************************************

Originally Posted by lemiant

FIRST is H-A-R-D!!! And to compete with the best your going to need comparable resources.

Let me fix that for you...

FIRST is H-A-R-D!!! And to beat the best often, your going to need comparable resources or a bit of luck.

Often times, the #1 alliance is the winner of the event. The #1 alliance (especially when it wins) is generally comprised with the 2 best robots in the division and then the 20th to 28th best (2nd round pick). If you want to repeatably beat the best, you must be around the top 4 at an event which is generally the area filled by the high resource teams. At many events, the 2nd round pick by the number 1 alliance is around the 50%-tile for the event. So, to compete with the best at a regional, you generally need to be top 20-ish of 40-60 teams, and for the world championship top 20-ish of 80-90 teams. This is easily achieveable for most teams with some organization and preparation, and delivering on relatively modest performance goals.

Per comments above, 27, 33, 1718 and approximately 300 (6-10 slots x 50+ events) others fall into a category of being really good but generally not the top 2-3 teams at any given event. When you are around the 3-10th best team at an event, you will more than likely fall into a valley of making elims, but not in an alliance strong enough to win. I would still consider this competing with those amazing teams, though not being able to beat them on most occasions. Within those 300 teams, you will find hundreds of examples of teams with significantly less resources maximizing their potential through benchmarking, smart design, hard work, and determination. Every year on Einstein, you will find teams that many might consider "lucky" for getting picked by 2 other awesome teams. More often, these teams have made their luck by performing really well and putting themselves into a position to "be lucky".

Notice, I mentioned 300 teams above. That leaves about 2100 other teams within FRC. Many of those are young teams some of which are over their heads and/or don't know what to do. A significant chunk though are teams that have the resources to be the 2nd round pick, but instead over-reach or are underprepared. There are a host of relatively young teams doing well year after year by coming up with good reasonable goals that challenge them, and then executing on those goals. While I do not know many of them outside of michigan, I can tell you 1718, 1918, 2054, 2137, 2337, 2612, 2619 2834, and 3098 have been steadily improving the last several years and have been beating many a vetran team. There are a large handful of other young teams that have shown a lot of promise but it takes more than 1 year to see how consistent they are going to be.

Of the 9 teams above, I could go on for probably an hour or more on how impressed I am with those teams executing their plans. The more informed will also note that the above teams not only are competitive on the field, but have also won business, website, Chairman's, and rookie-allstar awards. Some of them have even fostered a rookie team themselves (with the rookies being aprt of the handful I am watching/looking out for). These are teams that are operating at or near the highest levels of what FIRST is trying to achieve.

If you are only paying attention to the 12 teams making it to Einstein, then you are missing 99.5% (2400-12)/2400 of what FIRST is really about. It is really humbling to compare yourself to those guys. If you start focusing in on the success of the 300 I mentioned above you will find a lot of teams that regard their season successful without having to measure it against other teams success. You will find many improvements your team can make with little or no cost to your team towards becoming one of those 300 (hopefully this number will get larger).

Back on topic:
You will also find a lot of the 300 are teams with practice bots (who spent the $2-3K on making a practice bot instead of flying everyone to XYZ), or maybe they rebuilt their 2011 bot to be like one of the other teams they admire. Or maybe they made serious robustness improvements and competed at some off-season evnts for $100 or... maybe your team will talk with them at your next event and find some of the things they are doing right that you can do and then instead of debating the morality of in-equalities of FRC we can talk about how 4XXX learned a ton from team 2XXX in 2012 and is now kicking bot in 2013...

[Peter Matteson]

Originally Posted by lemiant

This is what I said. Rookies can be great, and I don't doubt that it is only a matter of time until one of them actually wins champs, but the odds ARE against them. Resources help you win, I don't think anyone here is going to challenge that.

And just for fun here's some math (Admittedly filled with all sorts of assumptions):

Last year, with numbers almost all the way to 4000, the un-weighted odds of a single team on the winning alliance being below 1000 was 25%... All three teams on the winning alliance were below 1000! The odds of that were 25%^3 or ~1.5%. that means that statistically the odds of all three teams on the wining alliance being below 1000 were just higher than the odds of any individual championship team winning. Obviously that is absurd! I could predict that the same will happen this year with at least a 50% chance of being right, you couldn't pick a winning team with anything approaching certainty.

Or take it one step further. 1114 was founded in 2003 so no team founded in a year after that would has ever won the championship. Assuming the ~2800 teams founded between 2003 and 2011 were founded in even increments of 350 teams per year the odds of none of these teams winning championships between 2004 and 2011 are:
((1200/1550) * (1200/1900) * (1200/2250) * (1200/2600) * (1200/2950) * (1200/3300) * (1200/3650) * (1200/4000)) ^3
or .00000054%.

I believe we can effectively determine from that number that veterans winning is not just statistical variation

Actually the odds are even worse than that...

Since FIRST went to the current divisional format in 2001 52.4% (74) of all Einstein slots (140 total) have been filled by just 24 teams with 2 or more Einstein trips!

35.7% of the total slots are taken up by just 12 teams with 3 or more trips.

42.9% (15) of all Championship Winner spots (35) in the same time frame were won by just 6 teams that have won 2 or more championships in divisional era!

The highest number team of the 24 is 1218.

The highest number team of the 12 is 968.

The highest number team of the 6 is 294.

All this said I'm willing to bet a coffee that with a list of 12 teams I can hit on 4 of the 12 Einstein competitors this year before I even know what anyone's robot looks like.

Edit: I forgot to cite my source, Jim Zondag's wonderful championship history white paper that he has published the last 2 years.