Ranking

[Collin Fultz]

Originally Posted by Mr. Lim

Pro-tip:

Friends don't let friends play defense! (in qualifications)

Good luck getting this to happen. Every match at KC has had defense played in it. Teams still want to win. And, judging by the scores across the country, KC isn't the only place this is true.

I am personally becomming a fan of this new ranking system. It's really factoring in a strength of schedule into the rankings. In the old system, a 2-0 win was worth the same as a 10-9 win (I know...ranking points...but the W was the most important). Now, a 10-9 win is a HUGE score.

We'll see how it plays out, but so far, I'm a fan. And, as I look through the top teams at KC on the rankings, I agree with most of them, which is more than I could say for the old system at times.

[Chuck Glick]

Originally Posted by Collin Fultz

I am personally becomming a fan of this new ranking system. It's really factoring in a strength of schedule into the rankings. In the old system, a 2-0 win was worth the same as a 10-9 win (I know...ranking points...but the W was the most important). Now, a 10-9 win is a HUGE score.

There is an easy way to factor in the difficulty of a match. Take both the red and blue scores (before penalties), add them together, and divide by 2 (or not). There is your match difficulty score. Teams are ranked by win/loss first, then difficulty score. This both promotes scoring and winning.

[Collin Fultz]

Originally Posted by Chuck Glick

There is an easy way to factor in the difficulty of a match. Take both the red and blue scores (before penalties), add them together, and divide by 2 (or not). There is your match difficulty score. Teams are ranked by win/loss first, then difficulty score. This both promotes scoring and winning.

Isn't this basically what the old system did? It ranked you by W/L then by opponent's points, so if you were winning high scoring matches, you ranked higher than if you were winning low scoring matches or matches where you scored high but your opponents don't score much.

I'm not saying one way is better than the other, yet. Let's get through a few weeks of regionals and Champs before hashing it out. All I'm saying is that I REALLY didn't like the ranking process going into this morning, but after watching it play out for a day, it's not so bad.

Will there be flaws in any system that uses such a small population of data to "rank" teams? Of course. That's the world in which we choose to compete. I'm just saying it's not as bad as I thought it would be going in.

[ubermeister]

I would be a fan of the new ranking system if the matches were higher-scoring. Here is an example of what happened to us:

We won a match 3-2 and were very happy, this was a fairly good match for us. Then, a robot on our alliance was assessed 5 <G46> penalties and the other team 1, causing us to lose 0-1. This means we get 2 seeding points. The other team gets 1 + 6 = 7 seeding points. This is a huge number of seeding points when the matches are usually won 1 or 2 to 0 (we only have around 8 seeding points total). A win due to exorbitant penalties on a good alliance is not a strategic win, and should not be rewarded like this.

[Vikesrock]

Originally Posted by ubermeister

A win due to exorbitant penalties on a good alliance is not a strategic win, and should not be rewarded like this.

If the alliance wracked up exorbitant penalties I guess they weren't really so good after all were they?

Moral of the story: Don't get penalties!!

[nikeairmancurry]

Yeah, but Michigan teams still need to win every match possible to quailfiy for states (State ranking points are based on wins/loses).. Kinda throws a monkey wrench into everything...

[Vikesrock]

Originally Posted by nikeairmancurry

Yeah, but Michigan teams still need to win every match possible to quailfiy for states (State ranking points are based on wins/loses).. Kinda throws a monkey wrench into everything...

While Michigan teams definitely have an added wrinkle, don't forget that making the Top 8 is also worth points in the Michigan system.

Being the #8 seed is worth more than 4 wins, and being the number 1 seed is worth the same as 8 wins.

[Smoores]

Yeah it seems that after 5 solid hours of live streaming today. A significant amount of teams don"t realize how they should play the game. Blocking hurts everyone, and multiple times people tried bumping robots trying to elevate. But holy cow, THERE WERE SO MANY PENALTIES!!!

[TD912]

Originally Posted by Smoores

THERE WERE SO MANY PENALTIES!!!

Yeah, I can't count the number of penalties there were for robots that somehow managed to drive on top of the soccer balls...

[dtengineering]

Well, as was posted somewhere else earlier.... "FIRST finally gives us a game where human beings can understand the scoring system, and then goes and gives us a ranking system that requires a two hour seminar on co-opertition to comprehend."

I get what FIRST is trying to do with the ranking system. I get the fact that Dean Kamen managed to get a patent on it... I've even read the patent (and the many objections the USPTO raised to it)... AND I get the fact that it is the rules, and that FIRST can set whatever rules they want. Even if I think the rule happens to be needlessly confusing, it is still a rule.

But it would really help to get people (who don't geek out over the FRC rules book) interested and excited about FRC if the ranking system could be quickly and easily understood by a person off the street without a lengthy lecture on "changing culture".

Thank goodness the elimination rounds will make sense to people who come to watch them.

Jason


P.S. My sympathies to those who don't "get it". It will all make sense if you watch a four hour retrospective video on the collected speeches of Dean and Woodie. You might still not LIKE it, but at least you'll "get" it.

[Wei]

FIRST has made winning or losing a match this year unimportant (for qualification matches). Its how everyone does (including your opponents) that affects your seeding score; not winning or losing. FIRST seem to have set up a version of the prisoner dilemma (game theory). Almost everyone I have talked to at the DC regional (on Friday) is playing this year's game as a zero sum game(my gain is your lost).

Imagine the scoring potential if all SIX robots were working together to score all the points for blue or for red. According to this years seeding formula:
winner seeding points = (winning alliance score - Penalty) + 2*(losing alliance score)
loser seeding points = winning alliance score

If everyone works together and the losing alliance does not have any points, then the winner and the loser get the exact same seeding score.

Where this is not a zero sum game is the number of balls score will be much higher when all six robots are working together then working against each other or the alliance leaving each other alone. Its very easy to move all the balls from the middle to the offense zone if there are 3 to 4 robots in the middle zone. You're more likely to have a robot free to handle a returning ball if the other robots taking a little longer dealing with their current ball(s). With the remaining 2 to 3 robot in the scoring zone, you have a situation where 1 to 2 robots are scoring while the other robot is collect balls to be score. With 6 working robots, the limitation on points will be how fast the human players can get the balls back into play, so that there are no penalties.

Under this strategy, the robots that cooperate the best together get the best seeding scores.

[Nawaid Ladak]

Originally Posted by Tarzan19

I understand the rules.
a day of competition made sure of that, it just baffled me that we were stomp and still shoot up 3 ranks, thats all.

the reason you shot up 3 spots was because nobody else played who was ranked above you, you pretty much gained nine points in the seeding match while everyone else didn't play.

I love this new rating system. I still do not comprehend why someone would want to play defense during qualifications, unless that's their outlet to get into the eliminations.

Speaking of that, im going to get a good laugh when those scores are much lower than what we're already seeing.

[sgreco]

I have declined to comment on the scoring system all year...

So all I have to say is: Any scoring system that ever gives people incentive to score on themselves at ANY point during ANY match needs to be re-evaluated.

What was wrong with the old days when the teams that won got the best seeds? What part of earning your seeding position isn't fair?

[Bongle]

Originally Posted by sgreco

I have declined to comment on the scoring system all year...

So all I have to say is: Any scoring system that ever gives people incentive to score on themselves at ANY point during ANY match needs to be re-evaluated.

What was wrong with the old days when the teams that won got the best seeds? What part of earning your seeding position isn't fair?

Unless, as posted above, FIRST's intent was to make a game where you aren't true opponents. Before the alliance era, there were games where (as I understand it) you had to work together with the others in the match to maximize your score in that match. If you look at backwards scoring less as "scoring on yourself" and more as "scoring for the greater good of both alliances", then the difference becomes clearer. This game and seeding system can be interpreted as not a strictly 3v3 game, it can also be seen at 6v0.

The highest possible QP comes from a tightly-fought match, but the most reliable QP comes if both teams agree to co-operate and only score on one side.

[JB987]

Originally Posted by Bongle

Unless, as posted above, FIRST's intent was to make a game where you aren't true opponents. Before the alliance era, there were games where (as I understand it) you had to work together with the others in the match to maximize your score in that match. If you look at backwards scoring less as "scoring on yourself" and more as "scoring for the greater good of both alliances", then the difference becomes clearer. This game and seeding system can be interpreted as not a strictly 3v3 game, it can also be seen at 6v0.

The highest possible QP comes from a tightly-fought match, but the most reliable QP comes if both teams agree to co-operate and only score on one side.

True if you could avoid penalties, which teams obviously are not doing