Selecting data by OPR

What is an OPR?

tip

TL;DR

OPRs are really complicated and hard to calculate. In the most simple of terms, it is an estimation of a team's individual contribution to the matches in which they competed. For those of us who don't quite understand the stats, the higher the better.

Bear with us, OPRs are very mathy.

OPR is now the more widely used term for what Karthik from Team 1114 once called "calculated contribution" — OPR attempts to calculate a team's contribution to their alliance's final score. If there were no alliances in FRC, a team's calculated contribution would simply be their average match score. This is an easy and straightforward metric to calculate. However, using average score with alliances, while giving you a positive predictor, isn't the most useful since the average score is a combination of the performance of all three teams on the alliance. 1114 starting using OPR in 2004 because they wanted a more useful metric to evaluate how teams had been playing than average score. They wanted a way to figure out, on average, what each team was contributing to their alliance's score. Here, OPR makes a key assumption: an alliance's final score is a linear combination of each alliance member's individual contribution. In other words, adding together the points earned by each member on a 3-team alliance results in the alliance's final score.

- The Blue Alliance

Types of OPRs

In addition to the simple average OPR, FTCStats also provides us OPRs for each phase of play, as well as a penalty-free OPR.

The non penalty, Auto, Teleop, and End Game OPR's are more specifically called OPRc's, which are similar to OPR but we are instead minimizing the Ax-b linear equation rather than solving for Ax=b in order to apply constraints such as ensuring scores >= 0 and imposing upper bounds on game-specific performance.

API Requests

Requests regarding OPRs are made using GET requests to the following endpoint:

https://ftcstats.jackcrane.rocks/api/performance/:type

The type parameter is required and can be one of the following:

• opr
• normalized
• non_penalty
• auto
• teleop
• end_game

The request should include an object in which you specify the values and logical operators you want to filter by:

{
  "operator": ">",
  "value": 0.5
}

Field	Type	Description
operator	string	The logical operator to use when comparing the value to the field. Refer to the following table for availible operators
value	number	The value to compare the field to.

Operators

Operator	MongoDB relation	Description
>	$gt	Greater than
>=	$gte	Greater than or equal to
<	$lt	Less than
<=	$lte	Less than or equal to
==	$eq	Equal to
!=	$ne	Not equal to

The API will return a JSON document with an array of PerformanceEntity objects:

{
  "count": 5,
  "entities": {
    [
      PerformanceEntity,
      PerformanceEntity,
      ...
    ]
  }
}

Javascript Example

Let's select all performances with an OPR greater than 154.

var myHeaders = new Headers();
myHeaders.append("Content-Type", "application/json");

var raw = JSON.stringify({ operator: ">", value: 154 });

var requestOptions = {
  method: "GET",
  headers: myHeaders,
  body: raw,
  redirect: "follow",
};

fetch("http://localhost:3001/api/performance/opr", requestOptions)
  .then((response) => response.text())
  .then((result) => console.log(result))
  .catch((error) => console.log("error", error));