Scouting History
Here's a chronology of the team's scouting history, starting from the most recent season:
2019 - Destination: Deep Space
Scouting Leads: Spencer Ng, ???
2018 - FIRST Power Up
Scouting Leads: Owais Aftab, Arushi Bhatia, Spencer Ng
This season saw the introduction of six Kindle Fire 7 tablets (with significantly better battery life) used for electronic scouting, with continuing use of the electronic scouting system from 2017, which involved the Android app, a desktop Java client, and Tableau for analysis. Assessments of scouts were done with a Google forms quiz, Kahoots, and Week Zero scouting, as well as scores for punctuality and reliability of scouts, which were all published in a public spreadsheet. These scores, along with several qualitative factors, were used to form the very-accurate scouting travel team, which used three rotations at the District Championship and two rotations at the World Championship.
As a time-based game, timers were used in parts of the scouting app, while solely using statistics to make judgement calls was deemphasized due to robots focusing on different aspects of the game depending on the match. Instead, many members of the scouting team took qualitative notes on robots' on-field capabilities and argued for or against them during the scouting meetings. Video review was heavily used to judge the speed and reliability of autonomous modes and power cube intakes, and we received a new camera with better battery life for the World Championship.
A new feature introduced this year was automatic picklist generation, based on ratings and pairwise-comparisons made by scouts with sequential robots that they scouted. These were fairly effective at providing a launch point for scouting meeting discussions, and adjustments were made with further pairwise-comparisons. Also introduced was a data auto-correction feature, which pulled from the Blue Alliance API to correct inaccurate data such as robot climbs and auto line crossings.
Our picking strategy involved forming an overall list of the best robots at district and offseason competitons, while forming distinct first/third and second picklists at the District and World Championships. We were fortunate to have compatible alliance captains approach us during four of our competitions this season due to our robot's outstanding on-field performance, which meant only needing to choose a second pick. We only emphasized quickly placing cubes on the exchange and switch at the beginning of the season due to the low skill floor, while having a climb and a (multicube) center switch auto was crucial toward the end of the season. For first picks, we emphasized a multicube scale auto and quickly and precisely placing cubes on the scale during tele-op.
Mount Olive District Event
- Accepted Team 303's offer as the first overall pick. Accepting their offer as the second seed was the best strategic decision, as the top-tier scale robots at the event were Teams 25, 303, and 1923.
- Selected Team 6203 as our second pick due to their exchange ability out of those in the remaining pool of robots.
- Event Winners
Montgomery District Event
- Accepted Team 225's offer as the first overall pick due to their three-cube scale autonomous and superior tele-op ability over the remaining pool of robots.
- Selected Team 5992 as our second pick due to their exchange ability out of those in the remaining pool of robots.
- Event Winners
Hatboro-Horsham District Event
- Accepted Team 2590's offer as the first overall pick due to their iron cross design that enabled an extra climb, their fast scale ability, and the lack of strategic options as the third seed. The top-tier scale robots at the event were Teams 25, 1218, 1640, and 2590, followed by Teams 1391 and 5407.
- Selected Team 6223 as the second pick due to their exchange ability out of those in the remaining pool of robots.
- Event Winners
FIRST Mid-Atlantic District Championship
- Selected Team 1640 for our first pick as the second alliance captain due to their two-cube scale autonomous, fast climb, and overall tele-op ability. Other options that we considered were Teams 365, 316, 1923, and 1712.
- Selected Team 1089 as our second pick due to their fast climb and center switch autonomous, as well as ability to play any role required, including the scale. We also strongly considered Team 2016 for their two-cube center switch autonomous and tele-op exchange, switch, and defense ability, despite lacking a climb and scale ability.
- Event Finalists
FIRST Championship: Detroit - Carson Division
- Accepted Team 225's offer as the first overall pick due to being the 19th seed. We were fortunate to be chosen despite our communication problems the previous day
- Selected Team 3015 as our second pick due to their climb, multicube center switch autonomous, and precise tele-op scale placement, which was crucial to our playoff strategy
- Selected Team 88 as our third pick due to their climb, exchange ability, and strategic calls made by their drive coach
- Division Semifinalists
2017 - FIRST Steamworks
Scouting Leads: Owais Aftab, Arushi Bhatia, Spencer Ng
This season saw a new set of scouting leads and a complete overhaul of the scouting subteam. Computer-based and practical assessments were given to the scouting team to form the basis of travel decisions, while the rules of the game were explained to scouts during meetings. This was highly successful and led to accurate scouting, partially due to the large number of freshmen scouts on the team. However, motivation for scouting tended to decrease as the season progressed.
The system itself was reinvented, with an Android app that converted collected data to JSON files that could be exported to the desktop client. The app itself utilized a modular, user-friendly material design interface with large buttons and checkboxes throughout, which could easily be modified for a given year. Meanwhile, the desktop client compiled the data from the six tablets together and provided team-based and alliance-based statistics, such as the overall average number of gears or the probability of one, two, or three climbs.
Tableau (introduced this year) and Excel were heavily used for data analysis, with Excel for looking at individual match data and Tableau for visualizing overall trends. In fact, much of the picklisting was done by looking at a plot of the average number of gears compared to takeoff percentage. The teams in the top-right indicated the best picks, while those in the bottom-left were the worst picks. We generally ignored autonomous capabilities due to the difficulty of a two-rotor auto and the reliability of our robot's autonomous mode. It is also notable that due to the "rope roulette" of the qualification matches, team rankings did not strongly correlate to robot ability.
Overall, despite our mediocore robot design, our scouting and strategy enabled us to select dark horses and win our first blue banners in five years.
Mount Olive District Event
- Selected Team 303 for our first pick as the second alliance captain due to their high number of gears, reliable climb, and shooter. They were the only objectively standout pick remaining after Teams 293 and 222 formed Alliance 1.
- Selected Team 1923 for our second pick due to their high number of gears, active gear placement, and reliable climb out of those available in the remaining pool.
- Event Winners
Montgomery District Event
- Accepted Team 75's offer as the eighth overall pick due to being 21st seed
- Selected Team 1923 for our second pick due to their high number of gears, center auto, and reliable climb out of those available in the remaining pool. This pairing enabled us to complete four rotors and up to two in auto.
- Event Winners
FIRST Mid-Atlantic District Championship
- Accepted Team 1923's offer as the eighth overall pick due to being the 37th seed
- Selected Team 75 for our second pick due to having the highest average number of gears, a gear intake, a side gear autonomous, and a reliable climb. We were also well-coordinated as an alliance after the Montgomery event, despite not having any fuel ability.
- Event Quarterfinalists
FIRST Championship: St. Louis - Carson Division
- Accepted Team 303's offer as the 17th overall pick (fourth member of the alliance) due to being the 25th seed
- Division Semifinalists
2016 - FIRST Stronghold
Scouting Leads: Arjun Bansil, Haley Williams
This year saw a new Android app, which collected data on defense crossings and boulders put into the high and low goals during both autonomous and teleop. The app exported a spreadsheet, which formed the basis of picklist discussions.
Due to our robot's outstanding low goal performance and reliable climb, we were fortunate enough to have picklist discussions with other teams before Alliance Selection, and our picklists tended to focus on the robots that could score the most points on average in a given match, through any means possible.
Hatboro-Horsham District Event
- Invited Team 341 for our first pick as the first alliance captain
- Selected Team 1807 as our second pick
- Event Semifinalists
Bridgewater-Raritan District Event
- Accepted Team 1257's offer as the first overall pick
- Selected Team 555 as our second pick
- Event Finalists
Mid-Atlantic Robotics District Championship
- Accepted Team 3314's offer as the third overall pick due to being the 11th seed
- Selected Team 1089 as our second pick
- Event Finalists
FIRST Championship - Curie Division
- Selected Team 2848 for our first pick as the third alliance captain. Many believe that Team 3641 should have been chosen instead.
- Selected Team 4329 as our second pick
- Selected Team 1885 as our third pick
- Division Quarterfinalists