Evaluating the Samsung Galaxy Watch8 Antioxidant Index: A 14-Day N of 1 Reliability and Dietary Trial
In the early 2000s I was introduced to the Pharmanex BioPhotonic Scanner, a handheld device that used resonance Raman spectroscopy to estimate skin carotenoids. At the time, I was genuinely impressed. It felt like a rare piece of consumer wellness technology that could provide objective feedback that you were actually eating enough colourful fruits and vegetables (Mayne et al., 2013).
Fast forward to today and the public-health problem has not gone away. In Australia, only 6.5% of people meet the vegetable recommendation, 44.1% meet the fruit recommendation, and only 4.2% meet both (Australian Bureau of Statistics, 2023). This is exactly the gap that makes a simple, repeatable feedback tool appealing. Even if it is not perfect, it can still help motivate behaviour change.
Samsung’s Galaxy Watch8 introduces an “Antioxidant Index” that is measured by pressing your thumb against the back sensor for about five seconds, aiming to estimate carotenoid levels in the skin (Samsung Newsroom, 2025a). The science behind the biomarker is credible. Skin carotenoid measurement using optical spectroscopy methods has been supported as a valid proxy for fruit and vegetable intake in the research literature (Mayne et al., 2013; Radtke et al., 2020). The key question is not whether skin carotenoids matter. The key question is how reliably and meaningfully a consumer wearable can detect change in real life. I decided to test it on myself.
How the measurement is taken
Figure 1. Galaxy Watch8 Antioxidant Index measurement method. The reading is taken by pressing the thumb firmly and steadily on the rear sensor for several seconds.
Why I Did This
I have not been consistently hitting the “5 + 2” target. Having a visible daily number was a strong nudge to tighten up my intake. This is the practical value of a feature like this. It can turn a vague intention, eat more vegetables, into something you can track and act on.
At the same time, I wanted to answer a basic measurement question: if my score changes, how much of that is real versus noise?
My two-week protocol
Reliability method (every morning on waking):
Same thumb, same general time, clean hands.
Three back-to-back measurements each morning.
Record yesterday’s vegetable and fruit serves.
Diet method (carotenoid-focused):
I structured daily intake around higher-carotenoid choices (tomato products, carrots, leafy greens, sweet potato, capsicum).
For practicality, I used V8 low-sodium vegetable juice as a daily back-up to help me reliably hit targets, acknowledging that juice is not a substitute for whole foods.
My Target carotenoid intake in milligrams
There is no formal daily “carotenoid mg” recommended intake, but skin-carotenoid methods tend to respond to sustained increases in carotenoid-rich foods over time, not single meals (Mayne et al., 2013; Radtke et al., 2020). My practical goal was to reliably reach a daily carotenoid exposure that was plausibly high enough to shift a skin-carotenoid signal.
A key anchor was one 250 mL V8 low-sodium serve daily, which is typically tomato-dominant and therefore likely to be lycopene-heavy, plus additional carotenoids from spinach, carrots, tomatoes, capsicum, mango, and sweet potato. This approach focused on foods that are realistically achievable day-to-day, not supplements.
What I Ate
Across the two weeks, my logged intake generally landed around 6 to 8 vegetable serves per day and 2 fruit serves per day, using a simple rotation of:
Spinach salads (2 cups), tomatoes or cherry tomatoes (150 to 180 g), carrots (75 g), capsicum (75 g),
Plus fruit (nectarine, peach, mango),
With V8 used on most days as a convenient back-up.
This matters because the watch feature is designed around carotenoids, so the best test is not “healthy eating” in general, but deliberately choosing foods high in the pigments the sensor is trying to detect.
My results in plain English
Here is the headline summary of my 14-day dataset:
Average daily score: 42.7
Day-to-day variability: CV 7.8%
Typical error (TE): 1.6 points
Smallest detectable change (SDC): 4.4 points
Example of a single reading and category
Figure 2. Example single reading from the Watch8 Antioxidant Index. This illustrates the score display and category label used by the app.
What SDC 4.4 Means
This is the most useful number.
It means that if my score changes by less than about 4 to 5 points, I should assume that change could simply be day-to-day measurement noise. If my score increases by more than about 4 to 5 points, I can be more confident that it reflects a real shift rather than normal variability.
So going forward, I have a practical rule:
Ignore small daily fluctuations.
Look for improvements of 5 points or more, ideally sustained as a weekly trend.
Did my score improve?
Over two weeks, my daily means ranged from the high 30s to the high 40s. Some of the higher values occurred during periods where my vegetable intake was consistently high. That said, this is still a short experiment and should be interpreted as a trend signal, not proof of causation.
Monthly trend view and range
Figure 3. Antioxidant Index trend view across the tracking period, showing the average score and the highest and lowest values, alongside the day-by-day pattern.
Where this fits, and where it does not
This feature is not a clinical measure of oxidative stress. It is not a diagnostic test, and it should not cause anxiety.
But as a behaviour tool, it has genuine potential. It did for me what the Pharmanex scanner did years ago. It made fruit and vegetable intake feel measurable and therefore harder to ignore.
It is also reasonable to acknowledge early public commentary that consumer implementations can behave inconsistently, and real-world use can reveal quirks that do not show up in development settings (Samsung Newsroom, 2025a; TechRadar, 2025; The Verge, 2025). That is exactly why establishing your personal SDC is useful. It helps you interpret the number like a measurement, not a judgement.
Take - home message
If you want to use the Galaxy Watch8 Antioxidant Index sensibly:
Measure consistently (same thumb, same routine).
Track weekly trends, not single readings.
Treat 5 points or more as a meaningful improvement threshold (based on my SDC of 4.4).
If you want the number to move, choose foods that actually contain the pigments being measured.
Use vegetable juice as a practical back-up if needed, but keep whole vegetables as the foundation.
For me, that is the value of this experiment. I now have a simple, repeatable feedback loop that supports a behaviour most Australians still struggle to sustain.
Appendix
Written By: Dr Luke Del Vecchio
Dany Williams
Typically replies within an hour
I will be back soon
Hey there 👋 It’s your friend Dany Williams. How can I help you?
Evaluating the Samsung Galaxy Watch8 Antioxidant Index: A 14-Day N of 1 Reliability and Dietary Trial
In the early 2000s I was introduced to the Pharmanex BioPhotonic Scanner, a handheld device that used resonance Raman spectroscopy to estimate skin carotenoids. At the time, I was genuinely impressed. It felt like a rare piece of consumer wellness technology that could provide objective feedback that you were actually eating enough colourful fruits and vegetables (Mayne et al., 2013).
Fast forward to today and the public-health problem has not gone away. In Australia, only 6.5% of people meet the vegetable recommendation, 44.1% meet the fruit recommendation, and only 4.2% meet both (Australian Bureau of Statistics, 2023). This is exactly the gap that makes a simple, repeatable feedback tool appealing. Even if it is not perfect, it can still help motivate behaviour change.
Samsung’s Galaxy Watch8 introduces an “Antioxidant Index” that is measured by pressing your thumb against the back sensor for about five seconds, aiming to estimate carotenoid levels in the skin (Samsung Newsroom, 2025a). The science behind the biomarker is credible. Skin carotenoid measurement using optical spectroscopy methods has been supported as a valid proxy for fruit and vegetable intake in the research literature (Mayne et al., 2013; Radtke et al., 2020). The key question is not whether skin carotenoids matter. The key question is how reliably and meaningfully a consumer wearable can detect change in real life. I decided to test it on myself.
How the measurement is taken
Why I Did This
I have not been consistently hitting the “5 + 2” target. Having a visible daily number was a strong nudge to tighten up my intake. This is the practical value of a feature like this. It can turn a vague intention, eat more vegetables, into something you can track and act on.
At the same time, I wanted to answer a basic measurement question: if my score changes, how much of that is real versus noise?
My two-week protocol
Reliability method (every morning on waking):
Diet method (carotenoid-focused):
My Target carotenoid intake in milligrams
There is no formal daily “carotenoid mg” recommended intake, but skin-carotenoid methods tend to respond to sustained increases in carotenoid-rich foods over time, not single meals (Mayne et al., 2013; Radtke et al., 2020). My practical goal was to reliably reach a daily carotenoid exposure that was plausibly high enough to shift a skin-carotenoid signal.
A key anchor was one 250 mL V8 low-sodium serve daily, which is typically tomato-dominant and therefore likely to be lycopene-heavy, plus additional carotenoids from spinach, carrots, tomatoes, capsicum, mango, and sweet potato. This approach focused on foods that are realistically achievable day-to-day, not supplements.
What I Ate
Across the two weeks, my logged intake generally landed around 6 to 8 vegetable serves per day and 2 fruit serves per day, using a simple rotation of:
This matters because the watch feature is designed around carotenoids, so the best test is not “healthy eating” in general, but deliberately choosing foods high in the pigments the sensor is trying to detect.
My results in plain English
Here is the headline summary of my 14-day dataset:
Example of a single reading and category
What SDC 4.4 Means
This is the most useful number.
It means that if my score changes by less than about 4 to 5 points, I should assume that change could simply be day-to-day measurement noise. If my score increases by more than about 4 to 5 points, I can be more confident that it reflects a real shift rather than normal variability.
So going forward, I have a practical rule:
Did my score improve?
Over two weeks, my daily means ranged from the high 30s to the high 40s. Some of the higher values occurred during periods where my vegetable intake was consistently high. That said, this is still a short experiment and should be interpreted as a trend signal, not proof of causation.
Monthly trend view and range
Where this fits, and where it does not
This feature is not a clinical measure of oxidative stress. It is not a diagnostic test, and it should not cause anxiety.
But as a behaviour tool, it has genuine potential. It did for me what the Pharmanex scanner did years ago. It made fruit and vegetable intake feel measurable and therefore harder to ignore.
It is also reasonable to acknowledge early public commentary that consumer implementations can behave inconsistently, and real-world use can reveal quirks that do not show up in development settings (Samsung Newsroom, 2025a; TechRadar, 2025; The Verge, 2025). That is exactly why establishing your personal SDC is useful. It helps you interpret the number like a measurement, not a judgement.
Take - home message
If you want to use the Galaxy Watch8 Antioxidant Index sensibly:
For me, that is the value of this experiment. I now have a simple, repeatable feedback loop that supports a behaviour most Australians still struggle to sustain.
Appendix
Written By: Dr Luke Del Vecchio