DEWBOT VI Scouting

Head Mentor(s):

1. Foster Shucker

Student Administrators:

1. Kenneth Au
2. DJ Raman
3. Ben Warner
4. Douglas Moyer

Mentor Moderators:

1. Foster Shucker
2. Clem McKown
3. Faith McKown

Introduction

For 2010 competitions, we wish to introduce novel scouting strategies and organization for better scouting and data collection.

Project Goals

1. Touch up ActiveCollab
2. Finish Camera System
3. Create an Access database after kick off
4. Develop new innovations for usage.

Old School

1640's current scouting system is significantly less technologically advanced (but easier to execute and at least approaching reasonably effective) than the proposed layout. The system involves two main aspects: pre-competition research and match observation.

Preliminary Research

Quantitative Metrics

• Team number

Rationale: Hard to scout without this.

• Team nickname

Rationale: Hard to scout without this too.

• Number of previous official competitions this season

Rationale: For many teams, the number of prior official competitions correlates strongly with robot performance. (Few teams do not improve over the season.)

• Week number and abbreviation of last official non-championship event, and of championship division if applicable

Rationale: Experience has shown that competitions prior to the last (two) official events provide little insight into current competition performance, especially compared to the amount of work required to generate the information. Knowing the week number is also helpful for this reason.

• Average score per match at last official non-championship event (2834)

Rationale: Together with winning margin, this separates the lucky-to-be-picked quarter-finalists from the unlucky-to-lose quarter-finalists, arguably even better than draft order.

• Average Winning Margin at last official non-championship event (2834)

Rationale: Works together with average score per match (above)

• Elimination Placings this season along with draft number for all official events

Rationale: This provides valuable insight (especially the draft numbers) into actual performance, and works well at a glance. Michigan District, Michigan State Championship, traditional Regional, Championship Division, and Einstein awards are all listed and tallied separately.

• Engineering and Other Awards of 1640 interest won this season, including multiplicities

Rationale: This is on a separate sheet from the standard database, but provides a good start for the "pit visit" list. "Of Interest" non-engineering awards currently include Chairman's, Engineering Inspiration, Woodie Flowers (with mentor name), Entrepreneurship, website, and Autodesk.

• Average Offensive Score, world percentile (1114)

Rationale: Great metric for identifying excellent strikers in the 2010 season.

• Calculated Contribution, world percentile (1114)

Rationale: Very helpful insight into overall robot abilities and contribution--this is who to watch out for.

• Seeding Score, world percentile (1114)

Rationale: Provides valuable comparison to calculated contribution. If these two differ considerably, the robot may not fit into the normal robot-type subdivisions.

• Average Defensive Score, world percentile (1114)

Rationale: Invaluable in determining who to watch out for. This is a great metric. Values of 90% and above are bolded for further recognition.

• Contribution to Winning Margin, world rank percentile (2834)

Rationale: This statistic was helpful at IRI given the sheer number of excellent teams present. The scouting/strategy teams will review its use for non-invitational events.

• Percent of Qualification Matches Won at last non-championship event, and championship division (in italics) if applicable

Rationale: This is a much more space-efficient way of communicating W-L-T statistics, and saves the arithmetic.

• Percentile Rank at last official non-championship event, and overall championship percentile rank (in italics) if applicable

Rationale: This helps put all the calculated numbers in perspective by relating them to a much more common metric.

• Hanging Points Scored at last official non-championship event

Rationale: This is an easy and relatively effective way to identify consistent and not-so-consistent hangers.

Past DEWBOTs Statistics for Reference

Many thanks to teams 1114 and 2384 for publishing their databases. 1114 in particular has published these for several years. These are the 1114 statistics for DEWBOTs IV-VI, for reference with other teams scouting numbers. All numbers are world rank percentiles.

DEWBOT & Event	Calculated Contribution	Average Offensive Score	Seeding Score	Average Defensive Score
DEWBOT IV - Hartford	21.3	11.0
DEWBOT IV - Philly	67.6	82.0
DEWBOT V - Chesapeake	10.3	22.5	77.4
DEWBOT VI - FLR	27.9	24.1	14.5	56.6
DEWBOT VI - Philly	75.3	80.8	90.3	31.4

Qualitative Notes

• Home, Mid & Away Zone Abilities (separately)

Rationale: These are graded 1-5, with 5 being the highest. Though the information is originally taken from watching minimal video archive footage, at least it's a place to start. "R1" indicates footage was only available from the first official event (omitted when scouting for an early on-season event), and thus should be assumed low.

• Kicker, Possessor & Defensive Abilities

Rationale: Never go on to the field without knowing as much as possible about opposing robots ball control, kicking, and defensive power. Coaching on field without it is asking for trouble.

• Autonomous Ability

Rationale: With any luck, this will give a hint about where the opposing balls will be at the end of autonomous, and where your alliance wants to set up for the match and/or immediately move to in teleop. For allied robots, it helps settle any "where should we start?" disputes.

• Notes

Rationale: These are used to explain the numerical rankings in the qualitative sections, as well as identify any other possible issues or incongruities.

Match Observation

Quantitative Metrics

• Starting Position

Rationale: This proves invaluable when determining where to set up alliance robots, and also provides an easy reference for balls scored in autonomous (which in turn provides a helpful correlation to kicker strength, if applicable)

• Scores in Autonomous

Rationale: This provides insight into both computerized aiming (most consistent high autonomous scorers have this) and where balls & the score will end up at the start of teleop. You've got to play the game differently if the chips are already down at the end of autonomous.

• Home Zone Misses in Autonomous

Rationale: This provides a reference point for autonomous scoring. You can't score if you don't shoot (and if you do shoot and don't make it, you've got other problems).

• Kicked Balls that Don't Clear the Bump in Autonomous

Rationale: This can provide invaluable insight into kicker/possessor strength, and is more consistent than teleop observations of the same phenomenon.

• Mid-to-Home, Far-to-Mid, and Far-to-Home Passes in Autonomous (separately)

Rationale: Great insight into kicker strength and accuracy, not to mention knowing where balls end up for teleop.

• Home, Mid, and Away Goals Scored in Telelop (separately)

Rationale: This is the point of the game, no? You can't play against (or with) what you don't know, and it's better to know beforehand.

• Mid-to-Home, Far-to-Mid, and Far-to-Home Passes in Teleop (separately)

Rationale: Provides a good comparison to the goals scored (above) as well as match strategy and where the bottlenecks may end up.

• Home Zone Misses in Teleop

Rationale: Much like in autonomous, this provides a reference point for striker accuracy.

• Kicked Balls that Don't Clear the Bump in Teleop

Rationale: Much like in autonomous and with Home Zone Misses, this reference point is important for the pass statistics and for kicker and possessor ability.

• Penalties

Rationale: Regular penalty-inducing robots are rare, but that makes them all the more important to know about

• Hanging (Y/N)

Rationale: Always good to know what robots are going to do in end-game, both so you know where they won't be and so you know what kind of cushion you do or don't need

Qualitative Notes

• Hanging - predicted consistency

Rationale: Currently, we can't watch every robot in every match. That makes predicting consistency, especially of a last-minute game-changing operation, very important.

• Ball Manipulation, Maneuverability, Defense Ability/Tactics, etc

Rationale: Things every driver and coach wants to know, and few manage to find out before the starting whistle.

Helpful Resources

US FIRST Team List

Information: Team number, official name/sponsors, nickname, location, rookie year, motto, website, and all official events and awards since 2001.

Details: Check "FIRST Robotics Competition" and select "2010 Breakaway". Un-check "Events" and "List Length (Max)" as desired. Hit "Show Results" and click "Team Number" to organize by team number. Alternatively, frclinks.com/t/

US FIRST Archived Events

Information: Team standings (rank, qualification* matches played for seeding, seeding score, coopertition bonus, hanging points), match-by-match results, and awards sorted by event for most official events since 2005 and the Championship since 2003. Helpful for getting perennial success information, especially with Week 1 Regionals.

Details: Scroll to bottom of page and select desired year and regionals and/or championship. Select regional or division's standings, match results, or awards. Alternatively, frclinks.com/e/[r for regional or c for championship]/[optional year] or frclinks.com/e/[optional matchresults, rankings, or awards]/[eventcode]/[optional year].

The Blue Alliance (TBA) Team Info

Information: Team number, official name/sponsors, nickname, location, rookie year, some robot photos, and match-by-match results and total* W-L-T from most official events since 2003.

Details: Enter team number in box. From a team page, change the URL to next desired number.

The Blue Alliance Match Archives

Information: Location, week number, and number of attending teams for official events back to 2003 the and 2002 Championship. Many match videos, though less in 2010 than other recent years.

Details: Select year and scroll to week and regional (check videos column for archive availability). Select regional and scroll to match; select film strip to view on page (YouTube style) or arrow to download into QuickTime (can be temperamental).

Chief Delphi (CD) Team Photo Search

Information: Provides team (robot, mechanism, anything uploaded to CD) photos and papers from all uploaded seasons. Also links to contibuter's usernames for further CD info.

Details: Select "photos" or "papers" and enter team number(s) or "frc[team number]", booleans accepted. Alternatively, frclinks.com/cdm/[team number]

"FRC Links" Navigating USFIRST.org

Information: Links directly to US FIRST, TBA, CD Media, pages (also Q&A and FIRST Forums) allowing for easy URL changing navigation.

Details: Follow usage instructions on home page

Team 1114's Pre-Championship Scouting Database (link is to CD white paper page for .xls file)

Information: Team number, team name/sponsors, nickname, rookie year, location, championship division (if registered). Also events attended; elimination finish, seeding score, qualification* W-L-T, seed, percentile, draft position, and awards for each pre-championship event. Also event-by-event average offensive, defensive, seeding and contribution score values, ranks, and percentiles relative to event and world (and rank among champions for winners).

FIRST Chat Scouting
Note: This site is not as thoroughly updated as FIRST, TBA, and 1114's.

Information: Team number, official name/sponsors, nickname, location, website, official events attending that season.

In "Robot Information": Drive type, steering, transmission, chassis orientation, manipulator type and capabilities, autonomous capabilities, and photo. Note that this is team-provided and open to public editing. The default in 2010 is 2wd, tank, 1-speed, long, with autonomous disabled.

In "Team Stats": qualification* W-L-T, seeding score, coopertition bonus, hanging bonus and match-by-match results for covered events. Balls scored, final position, OPR, CCWM, penalties, DQs, and disablings for scouted events.

Details:

Sorted by Team: Select team number range and then team number or enter number in the "jump to team" box. Select "Robot Information" or "Team Stats".

Sorted by Event: Select "Scouting" on the right-hand panel. Click the Regional or District drop down, type name or scroll to week and event as necessary. Click "Show Teams", click team number for team main page or "Stats" for team's stats that season.

FIRST Objective

Information: Official events attended this season, total matches, total W-L-T, average win/loss average (#Wins/#Matches), average winning and losing scores, overall average score, biggest win and loss margins, and highest-ranked alliance partners. Event analysis (total matches, highest score and alliance, average winning and losing scores) in beta.

Details: Enter team number or select event from menu and click "Analysis".

*Note that TBA and FIRST Objective list the total (qualification and elimination) W-L-T stats by event. FIRST, FIRST Chat, and 1114 include only qualification stats. Both methods have their benefits depending on the use, the season's game, and the seeding algorithm. Consider which you'd prefer, but be careful not to mix and match.

DEWBOT VI Databases

Pre-Competition Philadelphia Regional Database

The database used at the Philadelphia Regional on March 25-27 2010, including prior regional results and team information for attending teams, by Siri Maley. Does not include 1114's stats. Built from the US FIRST Team list, TBA team database, various CD forum post assessments, and previous 1640 competition experience.

Pre-Competition PARC XIII Database

The database used at the PARC XIII on May 8 2010, including prior regional results and team information for attending teams, by Siri Maley. Includes 1114's stats. Also built from the 1114's stats, US FIRST Team list, TBA team database, various CD forum post assessments, and previous 1640 competition experience.

Pre-Competition Monty Madness 2010 Database

The database used at the Monty Madness on May 15 2010, including prior regional and championship results and team information for attending teams, by Siri Maley. Includes 1114's stats. Also built US FIRST Team list, TBA team database, various CD forum post assessments, and previous 1640 competition experience.

There is no database for Bridgewater Battle.

IRI 2010 Database

The database used at the IRI Competition on July 16-17 2010, including prior regional and championship results and team information, allies, TBA videos, match observation, awards, etc. Prelim database by Siri Maley, match observation by Jen, Alex and Andrew. Includes 1114's and 2836's stats. Also built from US FIRST Team list, TBA team and video databases, and ChiefDelphi.

Pre-Competition Duel on the Delaware 2010 Database

The database used at the Duel on the Delaware on October 16 2010, including prior regional and championship results and team information, etc. By Siri Maley. Includes 1114's and 2836's stats. Also built US FIRST Team list, TBA team database, and ChiefDelphi.

Pre-Competition Ramp Riot 2010 Database

The database that will be used at Ramp Riot on November 13 2010, including prior regional and championship results and team information, etc. By Siri Maley. Includes 1114's and 2836's stats. Also built US FIRST Team list, TBA team database, and ChiefDelphi.

---