<!--
{
  "availability" : [
    "iOS: 9.0.0 -",
    "iPadOS: 9.0.0 -",
    "macCatalyst: 13.1.0 -",
    "macOS: 10.11.0 -",
    "tvOS: 9.0.0 -",
    "visionOS: 1.0.0 -"
  ],
  "documentType" : "symbol",
  "framework" : "GameplayKit",
  "identifier" : "/documentation/GameplayKit/GKGaussianDistribution/deviation",
  "metadataVersion" : "0.1.0",
  "role" : "Instance Property",
  "symbol" : {
    "kind" : "Instance Property",
    "modules" : [
      "GameplayKit"
    ],
    "preciseIdentifier" : "c:objc(cs)GKGaussianDistribution(py)deviation"
  },
  "title" : "deviation"
}
-->

# deviation

The standard deviation of the distribution (also called *sigma*).

```
var deviation: Float { get }
```

## Discussion

One deviation is defined such that 68.27% of values generated by the distribution are within one deviation of the [`mean`](/documentation/GameplayKit/GKGaussianDistribution/mean) value, 95% of generated values are within two deviations, and 100% of generated values are within three deviations. Values within two deviations of the mean have a 95% probabilty, and values within three deviations have a 100% probability.

> Note:
> In the standard mathematical model of a Gaussian distribution, 99.7% of generated values are within three deviations of the mean and the distribution’s range is infinite. GameplayKit discards the extremely unlikely values more than three deviations from the mean in order to set finite limits on the distribution’s range.

This property is read-only—GameplayKit automatically calculates its value based on those of the inherited [`lowestValue`](/documentation/GameplayKit/GKRandomDistribution/lowestValue) and [`highestValue`](/documentation/GameplayKit/GKRandomDistribution/highestValue) properties such that those values are each three deviations away from the mean (`deviation = (highest - lowest) / 6`).

---

Copyright &copy; 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)