# Epsilon

> The smallest positive non-zero value that a float can represent. Using Mathf.Epsilon as a delta for equality makes the comparison effectively identical to exact equality on normal-range numbers. For an approximately equal value, use Mathf.Approximately or a tolerance that reflects your data’s scale, typically a combination of absolute and relative tolerances.

## Definition

* **Type:** Field
* **Namespace:** [UnityEngine](/engine/6000.0/script-reference/unityengine.md)
* **Assembly:** UnityEngine.CoreModule

```csharp
public static readonly float Epsilon
```

### Remarks

Operations with Epsilon follow these rules:

* anyValue + Epsilon = anyValue

* anyValue - Epsilon = anyValue

* 0 + Epsilon = Epsilon

* 0 - Epsilon = -Epsilon

A value between any number and Epsilon will result in an arbitrary number due to loss of trailing digits.

### Examples

```csharp
using UnityEngine;

public class Example : MonoBehaviour
{
    // Avoid division by (near) zero using Mathf.Epsilon as a zero guard.
    float SafeDivide(float numerator, float denominator)
    {
        if (Mathf.Abs(denominator) <= Mathf.Epsilon)
            return 0f; // or handle appropriately

        return numerator / denominator;
    }
}
```