Poll-Video Replay

[gblake]

Ryan - Whether or not they are creating the right mental model(s), here is why I think most folks bring the big sports comparisons into the discussion.

They are following a thought process that takes them down this path:

1. Some people want to be able to answer questions about match events after-the-fact
2. The types of questions they want answered are similar to the questions big sports leagues try to answer (contact between players/objects, player/object location, when an event occurred, etc)
3. The player/object speeds, traffic jams, and chances for obstructed views are similar to big sports league scenarios
4. The action is less "interruptable" than most big league sports (this places burdens on the playback system's ease-of-use).
5. Because of the similarities between FRC and big league sports, getting the answers people want (see item #1) will probably require using a recording and playback set-up similar (in some respects) to the ones big league sports use.
   • Without an explanation of why they should think otherwise, I think that for most readers, a "drastically simplified" replay system sounds like something that would answer a drastically small subset of the questions people/teams want answered.

To get them thinking differently you need to interrupt that 5 step process, and insert something different that puts trains of thought onto a different set of tracks. You might insert statements like these:

• The system won't answer all questions. It will answer ___, ___, and ___.
• The system will have blind spots (X% of the total field, Y% of the high traffic areas of the field). Teams will know those spots in advance, and teams will know mistakes occurring in those spots can't be corrected using video replays.
• Depending on line-of-sight angles and distances, the system will determine locations with accuracies between +/-X and +/-Y. Teams will know this in advance and may plan gameplay and review-requests accordingly.
• A time-and-motion style analysis of the activities that will be involved in processing a review request predicts that processing each will add N seconds to the time between matches, so ...

Bottom line: In the absence of specific descriptions of why/how FRC questions are different from big sports questions, or need to be answered using different methods than the big sports methods, or are going to be limited to a subset of all the possible FRC questions, or are going to be answered well-enough (give metrics) using low-fidelity (give numbers) systems, or ... I think the "big sports" yardstick is going to be the one most people instinctively choose to use.

Blake

[ceb515]

There should be another option, "During Quarter-, Semi-, and Finals Matches"

[Wayne Doenges]

I'm missing something here. You mention wasting time. Do you think that EVERY team will ask for a video review of EACH match? I don't think that would happen.
If a team comes up and ask for a video review because they think there was a mistake made, the Head Ref should deny the request. If a team says that a crossing was missed and can give you an approximate time the crossing was made, than the Head Ref should consider the request.
We had one of the defenses come loose during a match (I won't say which match) and I feel it cost us the match. I could see it move every time a robot crossed it. That would also be a good time for a review.
But, just like in the NFL, you could maybe have a consequence for falsely asking for a review. I think this would limit to the ones that really occurred.
Just my $.05.
*puts on flame proof under garments

[christheman200]

After losing on a foul point tie breaker in quarters in GTRC, a video review would have been nice. Is it practical? I don't think so. The delays between matches were already long enough. However, if implemented it should be consistent from event to event, in other words fool proof. I like the idea of testing this at offseason events for a few years, and seeing where it goes.

[EricH]

Originally Posted by christheman200

However, if implemented it should be consistent from event to event, in other words fool proof.

I think something got lost in translation.

For one thing, "fool proof" doesn't exist. There's "fool resistant", but nothing is EVER fool proof. Anybody who thinks something is fool proof severely underestimates the ingenuity of the common fool (and the uncommon fool on top of that).

And for another, "fool proof" means that it's hard to mess up. Consistent from event to event means that it's the same in multiple places (within a certain degree of tolerance). Now, something that's consistent is likely to be MORE difficult to mess up the setup on, but they aren't the same thing. Example: If you want to be consistent, you use the same type of connector on all your wiring. If you want to be foolproof, you make sure that any given connector side can only go into it's matching one, and not into any others!


I think that for the initial rounds of testing (AKA, offseason events), variation--intentional variation--is good. This is called "exploring the alternatives". For example, will any old video work? That's one possible way to cut the costs down. Or camera variety A vs. camera variety B. Various locations can be checked (food for thought, if you've got a driver-cam, you probably can't tell if you actually got a Crossing on it unless it's a pole-mounted one).

After those initial rounds, everybody compares notes (publishes) and then the "standardization" work can begin. Run a couple rounds of checks on what's been shown to work decently, compare notes, repeat until you have a process.

[Tom Line]

Originally Posted by Ryan Dognaux

Here's what we use for our setup in case anyone needs to see a diagram of how this works. It's painfully easy to setup, spent maybe an hour setting it up by myself for St. Louis. We didn't even have a volunteer to run it, I would just occasionally check up on it. Having 1 dedicated volunteer or a couple to swap out means you can do fancier stuff, but it's really not required to sit at the laptop constantly all day. A field reset person could hang out there while taking a break just to make sure it still looks good.

We have banned the term "it's easy" from our team. It SHOULD be relatively easy to configure your laptop and radio to connect to the field... and yet that doesn't seem to be the case for many teams each year - even teams that already connected at a prior event. Put this in the hands of a hundred individuals at a hundred different regionals and watch total chaos commence.

Let me put this a different way. You would think mounting a go-pro on a pole and recording matches at an event would be fairly easy. Yet many regionals don't do it. Now take a step beyond that - it takes technical skills to turn that simple recording into a replay. Skills that 95% of folks that are volunteering at FIRST events don't have.

The funny thing here is that I'm arguing against something that I would actually like to have. We've had matches go for us, and against us, when they shouldn't have. I can show you a couple years ago where we should have moved on into the semifinals on our championship field but didn't because of a scoring error.

[Ryan Dognaux]

Originally Posted by Tom Line

Let me put this a different way. You would think mounting a go-pro on a pole and recording matches at an event would be fairly easy. Yet many regionals don't do it. Now take a step beyond that - it takes technical skills to turn that simple recording into a replay. Skills that 95% of folks that are volunteering at FIRST events don't have.

That's because a lot of regionals don't already own the equipment and view it as lower priority in the grand scheme of things. If the equipment traveled with the field and was part of the venue setup, I think you'd quickly see high quality video recording at every event (and in turn high quality webcasts assuming they have internet access.)

If we can set up a full size FRC field and get all of that working with FMS, surely each event can handle setting up an additional laptop, a camera on a pole and running a few connections cables. A bit more work? Sure. But c'mon, not really. A 2-3 page setup tip sheet is all that would be needed.

Also - it really doesn't take any additional skill to 'turn that simple recording into a replay' - you just stop live recording and go open up the video file that was saved.

Question to the volunteer - 'Can you open a file on a PC?' Congratulations you have the job!