Power Rankings Now Available

As a mentor with FRC team 1741, Red Alert, I have worked with the students performing the scouting and short list of teams for alliance selection. During that time I came across a scouting spreadsheet from team 2834 that we found very useful. The spreadsheet provided an Offensive Power Rating (OPR) and Calculated Contribution to Winning Margin (CCWM) for each team at each regional. You can find the scouting database and a presentation describing how the rankings are calculated at http://www.team2834.com/scouting_database.html. OPR and CCWM are calculated using only qualification matches, not elimination matches.

My goal was to provide the same type of information for VEX teams. I have the Power Rankings for Sack Attack to the point where they are ready to be made available for general viewing. I have calculated the OPR and CCWM for all teams at all competitions since the World Championship that have complete match results available on RobotEvents.com. I will continue to update the rankings as match results become available throughout the rest of the Sack Attack season. Some competitions have either incomplete or invalid match results, so I have not been able to perform calculations for those competitions. If you do not see information for a competition that you think should be available please let me know and I will see what I can do about getting the match results processed.

There are currently two ways to view the Power Rankings.

First, you can view the Power Rankings for all teams at a competition. For each team you can see their qualifying record, rank, WPs, SPs, tournament record, average scored, average opponent score, average winning margin, average losing margin, OPR, and CCWM. You can sort the results by clicking on the header of the column you with to sort by. (Note: The competition dropdown is not order all competitions correctly. I’ll take a look at that once I have made all the interesting data available.)

Second, you can view the Power Rankings from all competitions where a team has competed. The same information is available in this view, with one entry displayed for each competition the selected team attended. In addition there is a “Total” entry for the team. For qualifying record, tournament record, avg for, avg against, avg win, and avg loss the “Totals” entry does reflect totals for the entire Sack Attack season. For rank, WPs, SPs, OPR, and CCWM the “Totals” entry reflects the best value a team has achieved during the Sack Attack season.

I am working on a page to help with pre-competition scouting that will let you select a competition and see the results for all teams that will be at that competition. This will probably not be available until after the holidays.

If you have questions about what OPR and CCWM mean please read through the information in the link above thoroughly before asking questions. If you don’t see a competition that you think should be there or have problems accessing the Power Rankings, feel free to email me at coach@cornerstonerobitcs.net.

Finally, the Power Rankings can be found on the Cornerstone Robotics web site ([www.cornerstonerobotics.net) under the Power Rankings tab.

Enjoy.

Jay](www.cornerstonerobotics.net) under the Power Rankings tab.)

Great job. These rankings are very informative.

This was an interesting read, but I think the presence of SPs and the fact that most competitive teams pursue them makes both these systems unreliable for VEX competitions.

Both of the methods that teams use to artificially increase their SPs - scoring for their opponents and only descoring when the game is close - mean both systems can’t work as they should, I think. SPs invalidate the basic assumptions that teams that score more are of higher quality than teams who score less, and that teams who win by larger margins are of higher quality than teams who win by smaller margins. The OPR and CCWM systems are unable to distinguish between a point scored by your alliance against strong resistance and a point scored (or not descored when it could have been) by your opponents.

I could be wrong and it could be that there is a strong correlation between OPR or CCWM and other measures of team quality, but my initial reaction is that because of the SP issue this is unlikely to be the case.

I agree that this is far from a perfect ranking system, but as I said I still think it can be very useful for evaluating teams that that you are not familiar with or getting a high level view of teams that will be attending an upcoming competition.

Currently the OPR and CCWM are calculated only for qualifying matches. After the holidays I will be working on calculating OPR and CCWM for the elimination matches. This should remove much of the bias introduced by farming SPs, but will limit the data for fewer teams. Another issue with this is how some elimination data is reported back to VEX. At some competitions the elimination rounds always show the first two teams of the alliance playing. This would skew the results since it isn’t possible to know which two robots were actually on the field.

Jay

Yeah, the fact that a lot of organisers don’t record which alliances play in eliminations is a big data quality problem :/. I understand why it happens as eliminations are often the most hectic part of the event and people often want to get the games finished quickly rather than record data that might not ever get used for anything, but it definitely makes the robotevents eliminations data much less interesting and useful.

I can see how OPR and CCWM might be useful if you already know a team’s style of play. For example if a team has an undefeated record (so points scored for them aren’t an issue) and you know that they don’t play for SPs themselves, comparing their OPR to previous events would tell you whether they have improved or not.

Conversely, if you know nothing about a team then comparing their ranking to their OPR and CCWM might give you some idea of how offensive or defensive their play style is.

I think in terms of ranking teams from top to bottom, these tools are probably quite limited. But as part of a broader scouting strategy, they could be useful.

Absolutely! Your first line of data used should be from first hand experience with the teams. So much goes into scouting that can’t be easily quantified. What these might be able to do though is help you pick out a sleeper second pick that you may otherwise overlook. They could help you identify a team that went 2-4 in qualifying with their 2 wins being blowouts and their 4 losses all being by very small margins.

Right now you won’t be able to get OPR and CCWM during a competition. I still have to kick off the process manually. I hope by early next year to have it run automatically, calculating OPR and CCWM within a couple minutes of the final qualifying scores being available on RobotEvents.

Jay

lol. i also participate in frc, and i was thinking of doing this before worlds. But i guess i dont need to anymore. :slight_smile: