Teams to watch out for in week 3

[mwtidd]

Originally Posted by Donut

While I certainly love the publicity, take Arizona's practice match results with a grain of salt. There were so few robots practicing throughout the day, especially the early part (everything before match 15) that they were allowing team drivers to act as "robots" on the field to test the scoring system. Outside of 330 I don't believe anyone even got a minibot up the pole in practice.

Right.... This is why I only selected very specific teams, that looked as though they performed well for practice matches, in which most teams are dead. At the end of tomorrow I should have a much better picture of who the power players are. Also who some valuable underdogs may be. I've had a few team's ask for their specific contributions, and I'm developing a tier structure for all teams. So you can have an idea of how good a team is without even watching a match. That being said this isn't intended to replace real scouting, but rather compliment it. Create an objective metric to compare to subjective metrics.

[Lil' Lavery]

Might as well replace my scouts with a freshman with a twitter account.

[XaulZan11]

Originally Posted by Lil' Lavery

Might as well replace my scouts with a freshman with a twitter account.

That's what we plan to do. Not sure what I'll be doing with all my free time at Midwest...maybe wondering why we keep losing without good scouting info.

[Techhexium]

Originally Posted by Chris is me

That no one's mentioned 330 yet in this thread is ludicrous.

In addition, the absence of team 971 in this thread contributes to the Lunacy, but they don't compete until week 4...

Anyway I'm not too familiar with Sacramento teams but I think teams to watch out for can be 100, 604, and 3256.

[CassCity2081]

Just from the practice matches at Boilermaker today it looks like 1018 and 1501 are both heavy scorers, did not see a lot of minibots though.

I also have high hopes for 2081!

[Akash Rastogi]

Sacramento- 100, 604, 1323, 3256

[Dr Fritz]

ha ha yeah hopfully our mini bot points will creep up tomorrow

[martin417]

Originally Posted by ttldomination

Ah, modesty. I heard good things about 1771 as well. And no floor pick up my bahookey.

- Sunny

Not sure what a bahookey is , but for the record, we never claimed we didn't have floor pickup, only that it wasn't necessary to have a really good robot.

[XaulZan11]

Looks like 234 has a new fast minibot. They will not only be a team to watch this week, but at St. Louis with that new minibot.

[mwtidd]

Team 365 is dominating MD right now
15 pts robot ,16 pts minibot
probably inflating the the teams its been paired with too

Team 1484 in TX is worth watching
11 pt robot, 10 pt minibot

for boiler:
292
18 pt robot, 13 pt minibot
234
16 pt robot, 12 pt minibot

again the pts are fairly low because they are inflating their teammates, at the end of today I should have better pt values

[BrendanB]

Originally Posted by lineskier

Team 365 is dominating MD right now
15 pts robot ,16 pts minibot
probably inflating the the teams its been paired with too

Team 1484 in TX is worth watching
11 pt robot, 10 pt minibot

for boiler:
292
18 pt robot, 13 pt minibot
234
16 pt robot, 12 pt minibot

again the pts are fairly low because they are inflating their teammates, at the end of today I should have better pt values

How do you have a 16, 12, and 13 point minibot?

[mwtidd]

Originally Posted by BrendanB

How do you have a 16, 12, and 13 point minibot?

Thats their estimated value right now. I have an algorithm that pulls out:
ETC: estimated teams contribution
ERC: estimated robots contribution
EMC: estimate minibot contribution

These may be off right now, because of point inflation.
1114 at the end of qualifiers had a 33 pt robot and a 14 pt minibot,
meaning they alone were attributed to on average 47 points per match.

The end up being a good metric of both how good and how reliable a robot was.
I want to eliminate every teams 1 or 2 lowest scores, but the logistics of that are a little tough.

[thefro526]

I'll be very interested to see the Data once all of the Friday matches are over.

No offense, but using practice match data to rank teams is probably not the greatest system...

[AustinSchuh]

Originally Posted by Techhexium

In addition, the absence of team 971 in this thread contributes to the Lunacy, but they don't compete until week 4...

Anyway I'm not too familiar with Sacramento teams but I think teams to watch out for can be 100, 604, and 3256.

971 competes in week 5.

1868 is definitely a contender at Sacramento.

[mwtidd]

Originally Posted by thefro526

I'll be very interested to see the Data once all of the Friday matches are over.

No offense, but using practice match data to rank teams is probably not the greatest system...

Oh I didn't intend those rankings to be a *true* ranking of the teams.
I just wanted to show a few teams that were already putting up big #s in the practice matches. Putting up #s in practice matches is a very good sign, as most teams are still working out the kinks at that point. We'll see at the end of today how those teams I picked yesterday did today...

Also ETC only looks at qualifying matches, I was just curious if I used it to measure practice matches what the outcome would be. I also figured there wasn't much harm in saying check out a team. Where as come day 3 if you say a team is a good team, your metrics better be pretty certain of it.

I am working on a little algorithm to act as a alliance selection aid. Obviously accuracy would be much more important at this point than in the case of a heads up.

I think the #s are very important, especially after looking at FRCTop25, in which team 217 is very low ranked, where I would pick them over almost any other team (except 1114).

PR tends to swing the view of many people, having a good number to reference may make teams, that are less visible rise to the top.

Also as it differentiates between minibots and robots, it would help to pick a team thats a better match for your alliance. Again this would be used in collaboration with scouting information.