Health Score and Experience Index both measure how many active minutes of a user were affected by issues. This article uses one worked example for each score and explains why the results can differ even when they describe the same hour of user activity.
For direct comparison of the SysTrack Health Score and Experience Index values for your groups and systems, use the Health Score vs Experience Index DEX Pack.
How the Two Scores Differ
Health Score uses fixed health categories such as CPU, memory, disk, latency, startup, and faults. It measures how many active minutes were affected by one or more of those categories. When more than one category is active in the same minute, Health Score splits the weight of that minute evenly across the active categories:
Category impact cost = 1 / Number of active categories
Experience Index uses a curated set of sensors. Each sensor can contribute up to 5 eligible impact minutes in a 5-minute bucket. If the eligible impact across all triggered sensors exceeds 5 minutes, SysTrack allocates the 5 minutes proportionally among those sensors:
Sensor impact cost = Sensor eligible impact / Total eligible impact × 5
The two scores also organize time differently:
Health Score rolls impact into 10-minute active-time buckets and counts each affected minute once, even if more than one health category was active during that minute.
Experience Index uses 5-minute buckets and counts impact minutes for triggered sensors during active time. How often a sensor adds impact minutes depends on its Experience Index impact frequency. For the key concepts, see Understanding the Experience Index.
Health Score Example
This example uses 60 minutes of active time.
10-Minute Bucket | What Happened | Total Impacted Minutes |
|---|---|---|
08:00-08:10 | Startup affected 1 active minute | 1 |
08:10-08:20 | CPU affected 1 active minute and memory affected 1 non-overlapping active minute | 2 |
08:20-08:30 | Latency affected 1 active minute and CPU affected 2 active minutes, with 1 minute of overlap | 2 |
08:30-08:40 | Latency affected 8 active minutes and CPU affected 7 active minutes, with 5 minutes of overlap | 10 |
08:40-08:50 | No impact | 0 |
08:50-09:00 | No impact | 0 |
Explanation:
08:10-08:20 bucket: non-overlapping minutes add, so 1 + 1 = 2.
08:20-08:30 bucket: the overlapping minute splits evenly between Latency and CPU (0.5 each). The total cost of Latency for the bucket is 0.5, and the total cost of CPU is 1 (non-overlapping minute) + 0.5 (overlapping minute) = 1.5. Costs sum to 2, which matches the 2 impacted minutes in the bucket.
08:30-08:40 bucket: with 5 minutes of overlap for 2 categories, the bucket has 10 impacted minutes. Latency cost: 3+(5×0.5)=5.5 minutes. CPU cost: 2+(5×0.5)=4.5 minutes.
Total impacted time is 15 minutes.
Health Score = (1 - 15 / 60) × 100 = 75%
This example shows three Health Score rules:
Different affected minutes add together.
Overlapping affected minutes count once.
A 10-minute bucket cannot contribute more than 10 impacted minutes.
Experience Index Example
This example uses the same 60 minutes of active time.
Sensor impact minutes for this example:
Extended High CPU Use: 2
Low WiFi Signal Strength: 1
High Destination Latency: 2
Critical Application Hang: 3
Extended High CPU Use, Low WiFi Signal Strength, and High Destination Latency use the Once per scoring bucket while active impact frequency. Critical Application Hang uses Once per activation, so it adds its impact minutes only in the bucket where it triggers. This behavior depends on the Experience Index impact frequency of each sensor. For details, see Experience Index Impact Frequency Examples.
Assume the following Experience Index sensors triggered during the hour:
5-Minute Bucket | What Happened | Impact Cost (Minutes) |
|---|---|---|
08:00-08:05 | Extended High CPU Use triggered | 2 |
08:10-08:15 | Extended High CPU Use and Low WiFi Signal Strength triggered | 3 |
08:25-08:30 | Low WiFi Signal Strength triggered multiple times | 1 |
08:30-08:35 | Low WiFi Signal Strength stayed active, and High Destination Latency (2 impact minutes) and Critical Application Hang (3 impact minutes) triggered | 5 |
Explanation:
08:10-08:15 bucket: 2 + 1 = 3 impact minutes.
08:25-08:30 bucket: repeated triggers of the same sensor still contribute only 1 impact minute for the bucket.
08:30-08:35 bucket: Low WiFi Signal Strength contributes 1 eligible impact minute, High Destination Latency contributes 2, and Critical Application Hang contributes 3. The total is 6 minutes, so SysTrack allocates the 5 minutes proportionally: Low WiFi Signal Strength has an impact cost of 1/6×5=0.83 minutes, High Destination Latency has an impact cost of 2/6×5=1.67 minutes, and Critical Application Hang has an impact cost of 3/6×5=2.5 minutes.
Total Experience Index impact is 11 minutes.
Experience Index = (1 - 11 / 60) × 100 ≈ 82%
This example shows how Experience Index differs from Health Score:
Data is evaluated in 5-minute buckets instead of 10-minute periods.
If a sensor triggers multiple times in the same bucket, it contributes its impact only once.
If the eligible impact from triggered sensors exceeds 5 minutes, SysTrack allocates the 5 minutes proportionally among the sensors.
Experience Index Impact Frequency Examples
Experience Index impact frequency defines when an Experience Index sensor adds its impact minutes. For details, see Understanding the Experience Index. The following examples compare the two impact frequency values, Once per scoring bucket while active and Once per activation, for the same condition.
Sustained Condition Example
This example uses 60 minutes of active time. A sensor with 2 impact minutes is active for the whole hour, and no other sensor triggers.
Impact Frequency | What Happened | Impact Cost (Minutes) |
|---|---|---|
Once per scoring bucket while active | The sensor adds 2 impact minutes in each of the 12 buckets from 08:00 to 09:00 | 24 |
Once per activation | The sensor adds 2 impact minutes once, in the 08:00-08:05 bucket | 2 |
With Once per scoring bucket while active, total Experience Index impact is 24 minutes:
Experience Index = (1 - 24 / 60) × 100 = 60%
With Once per activation, total Experience Index impact is 2 minutes:
Experience Index = (1 - 2 / 60) × 100 ≈ 97%
Lookback Window Example
A sensor with a 10-minute lookback window checks for an event in the last 10 minutes. Because of this window, one event can keep the sensor active across several 5-minute buckets.
Assume that the sensor has 3 impact minutes and that one application hang occurs at 08:02. The sensor is active from 08:02 to 08:12, in three 5-minute buckets.
5-Minute Bucket | What Happened | Once per Scoring Bucket While Active (Minutes) | Once per Activation (Minutes) |
|---|---|---|---|
08:00-08:05 | The application hang occurs at 08:02, and the sensor triggers | 3 | 3 |
08:05-08:10 | The sensor stays active | 3 | 0 |
08:10-08:15 | The sensor stays active until 08:12 | 3 | 0 |
With Once per scoring bucket while active, the sensor adds 3 × 3 = 9 impact minutes. With Once per activation, the sensor adds 3 impact minutes, in the 08:00-08:05 bucket only.
These examples show how Experience Index impact frequency affects the score:
With Once per scoring bucket while active, the impact of a condition grows with the number of buckets in which the sensor is active.
With Once per activation, a condition adds impact minutes once, however long it lasts. If the condition ends and occurs again later, the sensor adds its impact minutes again.
The sensor window decides how long a sensor stays active, so one event can keep a sensor active across several buckets.
What to Expect When Comparing the Scores
In these examples, the same 60 active minutes produce two different results. This difference is expected. Health Score uses fixed health categories and counts overlapping categories within the same affected minute once. Experience Index uses named sensors, configurable impact minutes, and 5-minute bucket cost allocation.
If you compare both scores for the same user or device, do not expect them to match exactly. Instead, expect Experience Index to reflect a more specific set of issue detections and bucketed impact rules, while Health Score reflects impacted active time from the legacy health categories.