Amazfit Active 2 - Long-term Scientific Review

Product reviewed

Amazfit Active 2 (review title: “Long-term Scientific Review”). Price mentioned: ~$99 / €99.

Main features / what was tested

The reviewer evaluated performance using instrument-based references:

• Heart rate tracking, tested during:
  • Indoor cycling
  • Outdoor running
  • Outdoor cycling
  • Weightlifting
• Hardware coverage for testing:
  • Testing on two Active 2 units (first pre-production model; second production model)
  • A second tester (colleague Rafael) for some activities
• Sleep stage tracking
  • Reference: Zmax EEG headband (brain-wave based)
  • 8 nights analyzed
• GPS tracking
  • Cycling the same routes repeatedly
  • Checked signal overlap/consistency:
    • 4 rides home→work
    • 4 rides work→home

User experience / design impressions

• Described as “one of the prettiest” watches owned and very good-looking for the price.
• Feels solid.
• Reviewer liked the screen quality, calling it among the best for a smartwatch.
• Not sponsored; units were provided for testing, but the reviewer claims Amazfit had no influence on content.

Heart rate tracking performance (key findings)

Reference method

• ECG chest strap (Polar H10) used as the “gold-ish” reference for heart rate.

Overall conclusion on heart rate

• Heart rate accuracy is mixed to poor, with a recurring problem of detecting heart rate too low, especially in higher-intensity / dynamic activities.
• Retesting across:
  • a second unit
  • a second tester still showed largely the same behavior, suggesting it’s not just a single-unit fluke.

Indoor cycling

• Overall: “okay”
• Issue: sometimes tracks too low during parts of sessions
• Correlation: ~0.91
• Comparison context:
  • Better than some older Amazfit models (e.g., GTS4 noted as bad)
  • Worse than models like Amazfit Bip 6 / Amazfit T-Rex 3 (at least one performed noticeably better)

Outdoor running

• Worse than indoor cycling (noise/light/cadence effects)
• Correlation: ~0.81
• Error pattern:
  • Can drift ~10 BPM above or below actual heart rate for stretches (reviewer impression)
  • Multiple runs showed too low, too high, and even “bonkers” spikes
• Possible cause suggested: cadence lock / wrist rocking frequency confusing the optical sensor

Outdoor cycling

• Correlation: ~0.75 (significantly lower than indoor cycling)
• Recurring issue: many segments detect too low heart rate
• Additional problems: missing data and poor follow-through
• Not the absolute worst device overall, but clearly below better models (notably T-Rex 3; “Bip 6” had potential)

Weightlifting

• Reviewer: not good, especially at higher heart-rate peaks
• Tends to miss peaks (detecting too low during sets)
• Comparison:
  • Apple Watch is closer to “reliable,” but can still drop data
  • For the most reliable HR during lifting: use an external chest strap

Production vs pre-production unit (important detail)

• Reviewer worried about unit problems; Amazfit sent another unit.
• The “second” (production) unit showed similar patterns:
  • poor outdoors
  • low correlations
• Second tester (Rafael) results:
  • Running/cycling: mixed, but often showed drift and too-low readings
• Reviewer conclusion: the issue persists across testers and units.

Reported correlations (heart rate)

• Indoor cycling: ~0.91
• Outdoor running: ~0.81
• Outdoor cycling (first unit): ~0.75
• Outdoor cycling (second unit, larger dataset): ~0.59
  • Noted as 18 outdoor cycling sessions for that unit

Sleep stage tracking performance (key findings)

Reference method

• Zmax EEG reference
• 8 nights analyzed
• Review evaluates stage agreement percentages

Agreement by sleep stage (approximate values stated)

• Deep sleep: ~55% detected as deep
• Light sleep: ~80% agreement (some confusion with deep)
• REM (RAM) sleep: ~34% agreement (worst-performing stage)
  • Many REM segments predicted as light sleep (~63%)
• Awake time: ~34% agreement
  • Not emphasized because EEG detects many short non-real wake moments that skew statistics

Visual examples / qualitative notes

• Deep sleep overlap sometimes “not that bad,” but not perfect.
• REM/RAM tracking:
  • often too short, shifted, or missed
  • at least one example where the watch labels alternating segments as awake

Context vs other devices/brands

• Active 2 ranks lower-middle—not worst, but not competitive.
• Reviewer claims algorithm similarity with other Amazfit models:
  • T-Rex 3 and Bip 6 show similar sleep-stage algorithm behavior
  • Active 2 still isn’t great
• Reviewer’s sleep-stage “top picks”:
  • Best: 8 Sleep, Aura Ring, Sleep 2 (OURA?) / NOR app, Apple
  • Second tier: Whoop and Fitbit/Google Pixel Watch
• Conclusion: Active 2 is not recommended for sleep stage tracking, especially REM.

GPS tracking performance

Testing method

• Repeated rides on the same routes
• Checked overlap/consistency
• 8 total bike rides:
  • 4 home→work
  • 4 work→home

Overall finding

• GPS is the best-performing category compared with HR and sleep.
• Signal:
  • Acquired quickly
  • Decently consistent, though not perfect
• Reviewer rates it as “pretty decent” and better than expected, especially for the price.
• Comparison note: not as perfect as Garmin, but good enough for the price point.

Pros (unique positives mentioned)

• Excellent screen quality
• Very attractive / premium look for ~$99
• Solid feel
• GPS tracking is pretty good (fast lock and decent route consistency)

Cons (unique negatives mentioned)

• Heart rate tracking issues:
  • Often detects too low
  • Performance deteriorates with running and outdoor cycling
  • Possible cadence lock / wrist rocking interference
  • Weightlifting peaks are missed; tracking is unreliable
• Sleep stage tracking is weak, especially REM/RAM
• Overall: not reliable enough to depend on for HR-based training or detailed sleep-stage insights

Comparisons made (devices/models referenced)

• Amazfit T-Rex 3: generally better than Active 2 (HR + GPS context; sleep algorithm believed similar)
• Amazfit Bip 6: sometimes better than Active 2 (notably HR comparative ranking)
• Older Amazfit models: GTS4 and GTR4 mentioned as worse heart-rate performers
• Apple Watch: described as better for weightlifting HR reliability (though can drop data)
• Huawei Watch Fit (Fit 2/3/4 Pro): recommended as better for “decent HR tracking” on a budget
• Other strong sleep-stage performers: 8 Sleep, Aura Ring, Sleep 2/NOR, Whoop, Fitbit/Google/Pixel Watch, Apple

Verdict / overall recommendation (as stated by reviewer)

• If you care about heart rate tracking and sleep stage tracking (training + sleep insights): Do NOT recommend Amazfit Active 2 based on the reviewer’s results.

• If you mainly want a good-looking watch and reasonably good GPS, and you can accept unreliable HR/sleep-stage metrics (or plan to use an external chest strap for HR): It may be okay, especially at the ~$99 price point.

Unique points mentioned by speakers

• Single main speaker (Rob) throughout the review.
• Colleague Rafael contributes only via secondary testing results for heart rate (running + cycling), with no separate structured commentary beyond the data-based conclusions.