Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   Rules/Strategy (http://www.chiefdelphi.com/forums/forumdisplay.php?f=6)
-   -   Missed Inbounder Assists (http://www.chiefdelphi.com/forums/showthread.php?t=128350)

mathking 30-03-2014 23:03

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by magnets (Post 1366838)
I agree that this will be an issue. It's really tough for refs to see everything, and if a team notices that assists aren't being counted properly, what are the refs supposed to do? They aren't just going to trust what a random team member saw, and even if there is amazingly clear video evidence, they can't look at it. It's also not right to just say no to a team who says a match score is wrong. A replay might be a possible, but it ruins any surprise match strategy, and often doesn't give teams enough time to reset their robot.

Just out of curiosity, what's the policy on changing scores? Can a score be undone during a match? After a match? In eliminations/qualifications? Autonomous in teleop?

Referees can and do adjust scores after matches. Once the cycle has ended the change has to wait until the end of the game. There were a number of matches at Queen City in which we adjusted scores after matches. Sometimes we would hit the possession, score and cycle buttons in succession and realize the possession didn't take. We would note the cycle number and add the possession after the match was done. A few times we weren't sure if a team got a possession (mostly it was when they shot or passes just as they crossed from one zone to the next) so we entered the cycle and asked over the radio if there was a possession, fixing any discrepancy after the match.

The biggest problem is that the pads have a bit of lag, and you have to toggle between the foul and scoring/possession screens. Most of the time the pads worked well, but they wouldn't always register a button press. The time spent looking down to insure the button was pressed meant time not looking at the field. If we spend even more extra time trying to figure out if there was another possession or not then the new cycle wouldn't start promptly. I would guess about 10-12 times we went back and added possessions that added assists.

In most of those cases (maybe all of them) at least one of the teams came up to ask about it, and usually I would just turn and say something like "The third assist in your second cycle? Yeah we got it. We are adjusting the score." and they wouldn't even wait in the question box. I don't have an accurate accounting of how many assists we missed (I am sure we missed some) but there were probably a half dozen more times when teams came up and we said "Team xxxx did not get all the way into the white zone before they made their truss shot, so you didn't get three assists." Also, a couple of times when we asked each other "Did that shot go over the truss?" (there were a lot of high balls shot to human players that were above the posts and it was not always clear) the human player of the opposing alliance said "Yes it did."

As has been said in a few previous threads, it would probably have been better if the fifth referee had been included from the start and given a pad to enter fouls. I was the truss referee from Friday afternoon on, and it would have simplified things. But I highly recommend asking about missed possessions right after the match if you think you were shorted. There is a good chance that the referees will remember then and be able to correct it.

mathking 30-03-2014 23:12

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Karthik (Post 1366957)
Agreed. Which makes me think that perhaps it's time to go back to a game that can be scored at the end of the match with static field conditions. (Think 2007, 20091, 20112, 20133) Having the ability to recycle game objects into play opens up a wealth of options when it comes to game design, but until FIRST can prove that they can score matches with a much higher degree of accuracy, it might be time to scale back how these games are scored. This isn't a criticism, bu merely a simple understanding that there will always be scoring errors when scoring must be done in real time. Designing games to minimize the number of quick real time scoring decisions that need to be made will make things easier for the referees, and result in fewer complaints from teams.

1. Even with all scored balls being situated in trailers at the end of matches, there were still incorrectly scored matches. However, it was rare enough to be understandable.
2. Ignore the automated minibot scoring.
3. See #1, although these rare errors almost blew up on Einstein...

The flip side is when you don't have fairly accurate real time scoring people get really upset about not knowing the score as the game progresses. Just think about all of the complaints about the real time scoring last year.

After refereeing, I feel as though my first impression was correct. The problem is that we need at least two more people involved for scoring. If there were two more referees who were just tracking possessions (one for each alliance), and not ever having to toggle their screens, there would be few errors for possessions. Then the two near side referees could focus just on trusses, catches and scores. This would leave three referees (plus the head ref) to call fouls. Maybe you even call the first four scorekeepers (it might make it easier to recruit people) and the others referees.

I do think the problems will be much less at the Championships, when there are more experienced referees calling the games.

Lil' Lavery 30-03-2014 23:14

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Karthik (Post 1366957)
Agreed. Which makes me think that perhaps it's time to go back to a game that can be scored at the end of the match with static field conditions. (Think 2007, 20091, 20112, 20133) Having the ability to recycle game objects into play opens up a wealth of options when it comes to game design, but until FIRST can prove that they can score matches with a much higher degree of accuracy, it might be time to scale back how these games are scored. This isn't a criticism, bu merely a simple understanding that there will always be scoring errors when scoring must be done in real time. Designing games to minimize the number of quick real time scoring decisions that need to be made will make things easier for the referees, and result in fewer complaints from teams.

1. Even with all scored balls being situated in trailers at the end of matches, there were still incorrectly scored matches. However, it was rare enough to be understandable.
2. Ignore the automated minibot scoring.
3. See #1, although these rare errors almost blew up on Einstein...

Do you recall significant scoring errors in 2008, 2010, or 2012? I don't, but that doesn't mean they didn't occur. While the real time scoring displayed on the screen has been an issue for many games (including most that are scored post match), I think this year and 2006 are the only two years I remember a significant quantity of discrepancies in the final tallied scores.

Karthik 30-03-2014 23:33

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Lil' Lavery (Post 1366987)
Do you recall significant scoring errors in 2008, 2010, or 2012? I don't, but that doesn't mean they didn't occur. While the real time scoring displayed on the screen has been an issue for many games (including most that are scored post match), I think this year and 2006 are the only two years I remember a significant quantity of discrepancies in the final tallied scores.

2012 was fairly issue free as I recall. 2008 definitely had issues with missed hurdles, especially in those weird cases where balls didn't make a full lap. 2010 has issues in matches where DOGMA penalties came into play. Nothing nearly as bad as we've seen this year, probably a function of how overburdened the referees are this year compared to past years.

Karthik 30-03-2014 23:36

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by mathking (Post 1366985)
The flip side is when you don't have fairly accurate real time scoring people get really upset about not knowing the score as the game progresses. Just think about all of the complaints about the real time scoring last year.

This was not an issue in 2004, 2005, 2007, and 2011, which all had games where you could get a very good approximation of the score just by glancing at the field. That combined with some real time scoring for display purposes only, made all four of those games incredibly spectator friendly. To me, this type of setup is something we should be striving for.

RonnieS 31-03-2014 00:05

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Adam Freeman (Post 1366702)
**Disclaimer** Not trying to call anyone out...just trying to understand.

While reviewing our match footage from the Waterford District F3 match, I noticed that one of our balls was only credited for (1) assist, even though it was passed out of 4994 in the red zone to 51, who had a possesion all three zones.

https://www.youtube.com/watch?featur...pDfBkQ#t=12 2

We specifically worked with 4994 to modify their robot to be able to pass the ball quickly out of their machine via their roller intake. Ball ball went in and out of their robot very quickly, but it was actively ejected from the robot by the intake spinning.

In the end, it's a moot point since we lost 130-95, so regardless of the additional assist it would not have changed the outcome of the match. But, I know that the fact that the lights were only showing (1) assist effected our strategy of where 51 scored that ball (low goal v. high goal). At the time the score was very close, and the difference could have been scoring in the low goal and starting another cycle v. trying to continue to score in the high goal.

It might be hard to tell in the video, since there are robots blocking the inbound but I can assure you the ball passed through 4994's robot. Is there a specific amount of time refs are looking for before a ball is ejected? Do we need to review our strategy with the appropriate refs so they know what to look for when we are trying to get a third assist with this type of machine?

I was told this was not the only time we were not credited for this assist in the elims. Although, this is the only video example I could find. Possible it was just an honest mistake by the refs.

I'd like to utilize this strategy in future competitions, and I want to make sure we are doing it correctly, to maximize our score.

At livonia we had 2 matches in elims where the ref had missed 3 assists from 3 different cycles. They do have a lot to do but after it was brought to their attention the first time we thought it would be settled...they got it right the 3rd time though which is better than nothing. We have learned that sometimes you have to make these assists more clear than what we might think is a clear assist at the time in the match.

mathking 31-03-2014 00:44

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Karthik (Post 1366999)
This was not an issue in 2004, 2005, 2007, and 2011, which all had games where you could get a very good approximation of the score just by glancing at the field. That combined with some real time scoring for display purposes only, made all four of those games incredibly spectator friendly. To me, this type of setup is something we should be striving for.

I am not sure that I would equate "being able to glance at the field and approximate the score" with spectator friendly. In most sports and sports-like competitions you can't just glance at the field and see who is winning. You have to look at the scoreboard. I think accurate scoring is certainly something to strive for, and something we have a right to expect.

I don't think I would list 2004, 2005, 2007 and 2011 as good years for spectators. The games required some real explaining to understand. I remember a parent asking me "Why aren't they just picking up that white tube?" at the Championships in 2011. 2005 and 2011 were undoubtedly our two best years in terms of the quality of our robots performance they were not particularly spectator friendly games. And the minibot issues in 2011 pretty much disqualify it from any discussion of well done scoring in my mind. By the Championships 2004 was pretty much a race to see which side could get two robots to hang. 2010, 2012 and 2013 were all easier to understand and follow, endgames notwithstanding. Yes I agree that you could get a good approximation of the score by glancing at the field, but really that means you could tell if one side was obviously winning. I think that in 2012, 2013 and 2014 when one side is obviously winning it is pretty obvious to someone watching as well.

As for scoring errors, it is my general sense that there have been plenty of years where there were at least as many matches where the winners changed after the end of the match due to fouls and/or scoring changes. Last year there were certainly many more scoring changes after the end of the match. From 2010 - 2013 we have co-hosted and off season tournament, and I have been either the head referee, the announcer or the MC each year. So I have been on the sidelines watching a whole competition (not involved at all in coaching my team) and all of those years we had issues. In particular I feel as though this year a larger proportion of the penalties are being assessed as they happen. This is good, as it gives teams a better sense of the actual score and allows them to change strategy. It is bad because it gives the referee scorekeepers (since they are playing both roles) more to do and the system is not necessarily well set up to allow this.

PriyankP 31-03-2014 12:21

Re: Missed Inbounder Assists
 
I think we can all agree that no one is perfect. The referees are humans and they often make mistakes that doesn't affect anything in the grand scheme of things. But when a call goes against you, it feels like you have used cheated out of something.

I know a few people who volunteered their time to be referees and if there is one thing I know about them, it's that these individuals are smart and pay attention to the most minor details. Am I saying that they made no mistakes? No! I'm sure they contributed to the "inconsistent calls" but if they made such mistakes, I know for a fact that I would not be a perfect referee in their stead.


Instead of discussing things you can't predict (such as whether refs will be consistently incorrect or inconsistently correct) it maybe more beneficial to think of ways to help these individuals make correct decisions.

I quickly glanced over the thread and I saw someone mention how they informed the refs of pin possessions to get the assists. That is something that actually changed something. The refs were able to make a correct call because the action was brought to their attention.

If a ref misses something, bring it to their attention and they will not make that same mistake twice. If you are trying something novel, talk to the refs/head ref beforehand to make them aware of what they should look for!

Lil' Lavery 31-03-2014 13:24

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by Karthik (Post 1366997)
2012 was fairly issue free as I recall. 2008 definitely had issues with missed hurdles, especially in those weird cases where balls didn't make a full lap. 2010 has issues in matches where DOGMA penalties came into play. Nothing nearly as bad as we've seen this year, probably a function of how overburdened the referees are this year compared to past years.

I was considering 2010's DOGMA penalties to be a separate issue, but given that they were generated via the automated scoring system, I can see how they're related. I think the scoring system by itself was good at tracking score as far as I can recall. Once or twice there was an issue with multiple balls passing in quick succession, but I think that was an issue experienced less than once per event.

Quote:

Originally Posted by Karthik (Post 1366999)
This was not an issue in 2004, 2005, 2007, and 2011, which all had games where you could get a very good approximation of the score just by glancing at the field. That combined with some real time scoring for display purposes only, made all four of those games incredibly spectator friendly. To me, this type of setup is something we should be striving for.

For an individual on a team, especially ones who are well versed in the rules, I tend to agree (with the exception of some 2007 matches). For a layperson, I'm less convinced. 2004 and 2005 were staightforward enough that most could get a good handle on the score once they learned the point values. 2007 and 2011 I don't think were nearly as easy. While it would be obvious in a blowout, there were too many modifiers and too many endgame points to make it obvious who held the advantage in matches were both alliances were competitive. It was easy to see who held the longest row in 2007, but the exponential scoring didn't lend itself to quickly tabulating score for most spectators, and it would rarely be obvious if an alliance's lead was substantial enough to overcome the potential points of the end game. The diminishing returns on lower rows and the doubling value of ubertubes led to a similar issue with 2011.

Granted, nothing will ever be worse than 2001 for spectator friendly scoring. That game was a nightmare to try and explain and calculate scores.

waialua359 31-03-2014 14:57

Re: Missed Inbounder Assists
 
The DOGMA penalties were one of the worst ways teams could lose matches. I hope Human Players never get put into those situations ever again.:mad:

rich2202 31-03-2014 16:12

Re: Missed Inbounder Assists
 
I was an inspector 2 weeks ago, and asked a ref about when does bulldozing constitute possession? She said they had a standard, but could not disclose it. I'm guessing it is the same for pinning. I'll be Ref'ing this weekend, and will be interested in what they tell us.

From what I see in the video from a distance, I am not sure if I would have called it a possession, unless I was told in advance.. By the same token, I might not have called it a possession foul either. IMHO, you really have to hold the ball more than a fraction of a second. The ones I am more worried about are: Bot is running across the field, and hits the opponent's ball. When is that intentional?

In addition to telling the head ref before matches, you might want to hold the ball for 1 second prior to ejecting it. Or, only eject it immediately if it is a bot to bot pass (with no carpet inbetween). Otherwise, it just looks like it hit your bot and fell out (regardless of whether the motors were running or not).

pandamonium 31-03-2014 17:58

Re: Missed Inbounder Assists
 
Our team performed an assist strategy we developed now called "touch and go" I will link a video when it is uploaded. The basic concept is we have the ball in our roller intake resting on our bumper and we drive into our alliance partners. We do not release control of the ball but it meets the definition of trapped. We were consistently credited for this dozens of times. We had over 700 assist points because of this. We had one match where our alliance partners both played straight deffence and we drove around like crazzy running the ball into them and getting full assist points even though they did almost nothing other than deffence.

jsasaki 31-03-2014 18:40

Re: Missed Inbounder Assists
 
I hope they add more referees/scorekeepers at the division level in St. Louis. Although I've only used the assist/penalty panels (non HR panels) 2 or 3 times this year, I know how difficult it is to know exactly what assists went in while a match is in progress. As a HR standing in the middle of the field, the only things I could watch were truss score lights illuminating and high goals illuminating as assists occurred. Many times, I saw the third assist not lighting up because of possible FMS lag as soon as cycles ended. I hope FIRST can figure out a way to make assists more clear. There are a bunch of other things events can do to make things more clear, such as NOT rotating referees on qualification and elimination matches and letting us use more than the "allowed" amount of referees. Hopefully St. Louis will be better.

Siri 31-03-2014 18:47

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by rich2202 (Post 1367318)
I was an inspector 2 weeks ago, and asked a ref about when does bulldozing constitute possession? She said they had a standard, but could not disclose it. I'm guessing it is the same for pinning. I'll be Ref'ing this weekend, and will be interested in what they tell us.

From what I see in the video from a distance, I am not sure if I would have called it a possession, unless I was told in advance.. By the same token, I might not have called it a possession foul either. IMHO, you really have to hold the ball more than a fraction of a second. The ones I am more worried about are: Bot is running across the field, and hits the opponent's ball. When is that intentional?

Possession <a-b>, to include herding, have nothing at all to do with intention.
“carrying” (moving while supporting BALLS in or on the ROBOT or holding the BALL in or on the ROBOT),
“herding” (repeated pushing or bumping),

Technically <c-d> do reference non-physical criteria. Launch direction is "desired", though in practice I don't know of anything action that's met other G12c requirements and failed on "desired". Similarly trapping technically has a non-physical component, as it is "in an attempt to shield...", As A452 shows though, it hasn't been called purely that way, at least in recent weeks. (Side note: very frustrating, both for inconsistency in low goal possessions and the originally ruled-out viability of offensive traps like pandamonium's "touch and go", which we would've gladly designed for had it been described as legal during build.)
“launching” (impelling BALLS to a desired location or direction via a MECHANISM in motion relative to the ROBOT), or
“trapping” (overt isolation or holding one or more BALLS against a FIELD element or ROBOT in an attempt to shield them).

Intention also comes into play in deciding the foul now, though there's another inconsistency zone for intentional+inconsequential, depending on how it's read.
Violation: FOUL, if unintentional and inconsequential (i.e. does not significantly impact MATCH play). TECHNICAL FOUL per consequential instance. TECHNICAL FOUL per extended instance. If strategic, RED CARD for the ALLIANCE.

Finally, intention is associated with advertency of bulldozing. (More inconsistency on whether these are identical--can one be attentive and advertent that they're bulldozing without it being intentional?) This is a lesser issue though, because deflecting has no positive or negative intention requirement.
B. “deflecting” (a single hit to or being hit by a BALL that bounces or rolls off the ROBOT or a BALL slips through the grips of a ROBOT without arresting the BALL'S momentum).

What you're describing as being worried about is deflection, which is not possession regardless of intent.


Re: 3rd assist high goal lights. Yes, the lag is terrible between the panels and the lights. You've really got to go by score. I've also seen it miss a last-minute possession entered by the far side ref while the near side hits end cycle. We have to fix this post-match.

engunneer 01-04-2014 06:52

Re: Missed Inbounder Assists
 
Quote:

Originally Posted by mathking (Post 1366978)
Referees can and do adjust scores after matches. Once the cycle has ended the change has to wait until the end of the game. There were a number of matches at Queen City in which we adjusted scores after matches. Sometimes we would hit the possession, score and cycle buttons in succession and realize the possession didn't take. We would note the cycle number and add the possession after the match was done. A few times we weren't sure if a team got a possession (mostly it was when they shot or passes just as they crossed from one zone to the next) so we entered the cycle and asked over the radio if there was a possession, fixing any discrepancy after the match.

The biggest problem is that the pads have a bit of lag, and you have to toggle between the foul and scoring/possession screens. Most of the time the pads worked well, but they wouldn't always register a button press. The time spent looking down to insure the button was pressed meant time not looking at the field. If we spend even more extra time trying to figure out if there was another possession or not then the new cycle wouldn't start promptly. I would guess about 10-12 times we went back and added possessions that added assists.

In most of those cases (maybe all of them) at least one of the teams came up to ask about it, and usually I would just turn and say something like "The third assist in your second cycle? Yeah we got it. We are adjusting the score." and they wouldn't even wait in the question box. I don't have an accurate accounting of how many assists we missed (I am sure we missed some) but there were probably a half dozen more times when teams came up and we said "Team xxxx did not get all the way into the white zone before they made their truss shot, so you didn't get three assists." Also, a couple of times when we asked each other "Did that shot go over the truss?" (there were a lot of high balls shot to human players that were above the posts and it was not always clear) the human player of the opposing alliance said "Yes it did."

As has been said in a few previous threads, it would probably have been better if the fifth referee had been included from the start and given a pad to enter fouls. I was the truss referee from Friday afternoon on, and it would have simplified things. But I highly recommend asking about missed possessions right after the match if you think you were shorted. There is a good chance that the referees will remember then and be able to correct it.

As a controls engineer who designs HMIs for machine control, this seems very much like a UI problem on the panels. I haven't used AB stuff (The pads are all Allen Bradley), but just about any plc will have trouble with very fast button presses, and will have a page change time. Any HMI design should try to minimize quick succession button presses and page changes. Humans can be way faster during something like an FRC event once they know where the buttons are, compared to the HMI designer who tested the panel.

Another thing is that the update rate of the HMI software may just not be fast enough. You can get a big usability bump from changing the screen -> plc communications rate from a typical 100 or 200ms down to 25 or 50.

I'd be interested in a list of the buttons on the possesion/foul screens, to see if I could work out a single page, minimal press version.


All times are GMT -5. The time now is 09:13.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi