<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
<title>DuckDik Blog</title>
<link>https://duckdik.com/blog/en/</link>
<atom:link href="https://duckdik.com/blog/en/feed.xml" rel="self" type="application/rss+xml"/>
<description>Informative articles on what wearables measure, what the numbers mean, the latest advances in sensors and how artificial intelligence helps make sense of sleep, recovery and everyday health. No jargon.</description>
<language>en</language>
<lastBuildDate>Mon, 16 Nov 2026 10:29:48 +0000</lastBuildDate>
<image><url>https://duckdik.com/assets/logo.png</url><title>DuckDik Blog</title><link>https://duckdik.com/blog/en/</link></image>
<item>
<title>How a screenless wristband counts your steps</title>
<link>https://duckdik.com/blog/en/how-a-screenless-wristband-counts-your-steps/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/how-a-screenless-wristband-counts-your-steps/</guid>
<pubDate>Mon, 16 Nov 2026 08:00:00 +0000</pubDate>
<description>A silicon comb smaller than a lentil turns the swing of your arm into steps. Here is how it works, why it slips up and where 10,000 steps came from.</description>
<content:encoded><![CDATA[<h2 id="a-sensor-smaller-than-a-lentil">A sensor smaller than a lentil</h2>
<p>Inside a screenless wristband there is a tiny object that works non-stop, 24 hours a day: the accelerometer. It is smaller than a lentil, and yet it knows at every moment which way your arm is moving and how hard. It is the source of your daily step count, much of the information about your sleep phases and the clue that tells whether you are sitting still or walking.</p>
<p>There are no wheels or pendulums, as in the pedometers of decades ago. Instead there is a comb-shaped silicon structure, with teeth thinner than a hair, suspended on microscopic springs. When you move your wrist, this piece shifts by a fraction of a micron through sheer inertia, just as your body lurches forward when the bus brakes. As it moves, the distance between the teeth and a set of fixed plates changes, and with it a tiny electrical charge. The chip translates that change into an acceleration figure.</p>
<h2 id="three-axes-and-gravity-always-there">Three axes and gravity always there</h2>
<p>The sensor does not measure in just one direction but in three: forwards and backwards, side to side, and up and down. And it always feels a force, even when you are perfectly still: gravity. That also lets it work out how your arm is oriented. Whether your wrist is lying face up on the table or hanging beside your hip, gravity “falls” on a different axis.</p>
<p>To avoid draining the battery, the sensor takes a few dozen samples per second, and a small processor summarises the data before sending it to the phone. That is one of the reasons a screenless wristband can last for weeks on a single charge.</p>
<h2 id="how-a-swing-becomes-a-step">How a swing becomes a step</h2>
<p>When you walk, every footfall produces a small jolt and your arm swings with a very regular rhythm. In the accelerometer’s signal this shows up as a series of peaks, almost like waves breaking on the shore. The algorithm does three things: it smooths the signal to remove tremors, looks for peaks that rise above a certain threshold and checks that they arrive at a believable cadence, roughly between one and three steps per second.</p>
<p>That last filter is the key. A single sudden movement does not count. Before the wristband starts adding up, it needs to see several peaks in a row with a steady rhythm. That is why the first steps after you get up from your chair sometimes don’t appear until you have been walking for a few seconds: the system was waiting to confirm that you really were walking.</p>
<h2 id="phantom-steps-and-lost-steps">Phantom steps and lost steps</h2>
<p>No counter is perfect. Some movements look a lot like walking: clapping, whisking eggs, brushing your teeth or driving along a bumpy road. The algorithms try to discard them, but a few slip through.</p>
<p>The opposite happens too. If you push a shopping trolley or walk with your hands in your pockets, your arm barely swings and the wristband may miss some of your steps.</p>
<p>That is why your wristband and your phone almost never agree. The phone, in your pocket, feels the jolt of your hip; the wristband, the swing of your arm. Neither is lying: each one sees the same walk from a different place. What is useful is to compare each device with itself, day after day.</p>
<h2 id="where-10000-steps-came-from">Where 10,000 steps came from</h2>
<p>So where does the famous daily target come from? Japan, in 1965. The year after the Tokyo Olympics, the company Yamasa launched the first consumer pedometer, the “manpo-meter”, which means something like “10,000-step meter”. It got that name because it was riding a popular campaign that encouraged people to walk ten thousand steps a day: about seven kilometres with 70-centimetre strides. It did not come out of any laboratory.</p>
<p>Science has taken decades to put numbers on it. A large analysis published in 2025 in The Lancet Public Health, which pooled 57 studies, found that going from 2,000 to 7,000 steps a day is associated with a 47% lower risk of death from any cause. Beyond that point the benefit continues, but more and more slowly. Ten thousand is still a good goal for people who are already active, but it is no magic line.</p>
<p>Try this experiment this week: count a hundred steps under your breath during a walk and then check how many your wristband has added up. You will find out how accurate your counter is and, while you are at it, how much your arm swings when you walk.</p>]]></content:encoded>
<category>steps</category><category>accelerometer</category><category>sensors</category><category>how it works</category>
<enclosure url="https://duckdik.com/blog/img/2026-11/pulsera-cuenta-pasos-acelerometro-8e1edcc2.webp" type="image/webp" length="0"/>
</item>
<item>
<title>The siesta isn&#x27;t as Spanish as we think, nor does it cost us as much sleep</title>
<link>https://duckdik.com/blog/en/siesta-europe-data-2026/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/siesta-europe-data-2026/</guid>
<pubDate>Tue, 10 Nov 2026 08:00:00 +0000</pubDate>
<description>One in five takes a weekly siesta, Germans almost as often as Spaniards, and a long nap barely shortens the night: data from 4,289 siestas in five countries.</description>
<content:encoded><![CDATA[<p>The siesta has a reputation for being Spanish, but the data tell a different story. Germans nap almost as much as Spaniards, the Portuguese take the longest breaks, and Sunday is Europeans' favourite day for a snooze.</p>
<p>These are some of the findings of a study by DuckDik, the Spanish wellness wristband company, which analysed <strong>4,289 siestas recorded by 1,205 users</strong> in Spain, Italy, France, Germany and Portugal over the past 60 days.</p>
<p>The results challenge some clichés. Seven in ten participants took at least one siesta during the two months analysed, and one in five naps at least once a week.</p>
<p>Spain tops the ranking, with 23% of users taking a siesta every week, but the gaps with other countries are smaller than one might expect. Italy reaches 20%, Germany 19% and France 18%. Portugal, at 11%, is the country where the habit is least common, although its siestas are the longest.</p>
<p>“We've always thought of the siesta as a very Spanish habit, but it turns out Germans take almost as many as we do. Perhaps the siesta is far more European than we imagined,” says Alberto Lorente, co-founder and CEO of DuckDik.</p>
<h2 id="sunday-is-the-king-of-the-siesta">Sunday is the king of the siesta</h2>
<p>The typical siesta lasts 40 minutes, although many people take their time. 38% go beyond 45 minutes, and one in ten lasts more than an hour and a half.</p>
<p>Italians and the French take the shortest ones, with a typical length of 35 to 36 minutes. The Portuguese, although they nap less often, are the ones who stretch their siestas out the most.</p>
<p>The day of the week also plays a role. <strong>Sunday is the big siesta day:</strong> one is recorded on 12% of the Sundays analysed, compared with around 8% of weekdays.</p>
<p>In Italy, siesta frequency almost doubles at the weekend. In Spain, by contrast, the difference is smaller, because Spaniards are also the ones who most often turn to this break during the week.</p>
<p>Age makes another difference. From the age of 65, one in three users takes a siesta every week, although their naps tend to be shorter. Under-35s, on the other hand, take the longest siestas.</p>
<h2 id="what-if-the-siesta-didnt-cost-us-so-much-sleep-at-night">What if the siesta didn't cost us so much sleep at night?</h2>
<p>One of the study's most interesting findings concerns a common worry: the fear that sleeping in the afternoon will stop you from resting well at night.</p>
<p>To look into this, DuckDik compared each user's nights after a siesta of more than 45 minutes with their own nights without a siesta, taking the same day of the week into account.</p>
<p>The result was striking: <strong>after a long siesta, participants slept on average just six minutes less at night</strong>. Bedtime barely changed.</p>
<p>In Spain the difference was somewhat larger, at around 13 minutes less night-time sleep.</p>
<p>This does not mean that a long siesta never harms rest, but it does challenge the idea that sleeping in the afternoon necessarily ruins our night. Among the users analysed, the average reduction in night-time sleep was small.</p>
<p>“Lots of people avoid the siesta because they think they won't sleep afterwards. The curious thing is that our data show a difference of just a few minutes. Everyone is different, and what matters is finding out how your own body responds,” explains Lorente.</p>
<h2 id="two-months-of-siestas-in-five-european-countries">Two months of siestas in five European countries</h2>
<p>The study is based on aggregated data from 1,205 DuckDik wristband users between 9 September and 8 November 2026. In total, 4,289 siestas lasting at least 20 minutes were analysed.</p>
<p>This is an observational analysis that describes participants' habits, without claiming to represent the European population as a whole or to establish cause-and-effect relationships. To protect privacy, only results for groups of at least 50 people are published.</p>
<p>The full report and methodology are in the annex below and in the <a href="https://duckdik.com/en/prensa/notas/siesta-europa-2026/">press release</a>.</p>
<h2 id="annex-data-by-country-and-methodology">Annex: data by country and methodology</h2>
<table>
<thead>
<tr>
<th>Country</th>
<th>People</th>
<th>Siesta at least once a week</th>
<th>Any siesta in 60 days</th>
<th>Typical duration</th>
</tr>
</thead>
<tbody>
<tr>
<td>Spain</td>
<td>459</td>
<td>23%</td>
<td>75%</td>
<td>40 min</td>
</tr>
<tr>
<td>Italy</td>
<td>286</td>
<td>20%</td>
<td>70%</td>
<td>36 min</td>
</tr>
<tr>
<td>Germany</td>
<td>258</td>
<td>19%</td>
<td>67%</td>
<td>40 min</td>
</tr>
<tr>
<td>France</td>
<td>112</td>
<td>18%</td>
<td>65%</td>
<td>35 min</td>
</tr>
<tr>
<td>Portugal</td>
<td>90</td>
<td>11%</td>
<td>64%</td>
<td>43 min</td>
</tr>
</tbody>
</table>
<p><strong>Days with a siesta by day of the week:</strong> between 7.5% and 8.1% from Monday to Friday, 9.7% on Saturday and 12.0% on Sunday. Weekdays vs weekend: Spain from 9.1% to 11.7%, Italy from 7.0% to 11.5%, Germany from 7.8% to 9.8% and France from 5.9% to 8.4%.</p>
<p><strong>Siesta and the following night:</strong> for each person, the night after a siesta of more than 45 minutes is compared with their own nights without a siesta on the same day of the week (Sunday with Sunday, Tuesday with Tuesday), across 497 people in the five countries.</p>
<p><strong>Methodology:</strong> daily aggregated data that the DuckDik app syncs with the user's account (never raw signals), from 9 September to 8 November 2026. People with at least 21 days of records are included; each person's country is the one that appears most often for them. “Weekly siesta” means at least one siesta for every seven days with records. Only siestas of 20 minutes or more that the app recognises or that the person confirms are counted. Figures are given only for groups of 50 people or more; the over-65 group comprises 79 people. Valid nights: between 3 and 14 hours of sleep, with a bedtime between 19:00 and 06:00. Observational study: it describes this sample of DuckDik users, not the whole population.</p>]]></content:encoded>
<category>sleep</category><category>siesta</category><category>data</category>
<enclosure url="https://duckdik.com/blog/img/2026-11/siesta-europa-2026-cover-9764c2cf.webp" type="image/webp" length="0"/>
</item>
<item>
<title>How your wristband knows you are stressed without asking</title>
<link>https://duckdik.com/blog/en/how-your-wristband-knows-you-are-stressed/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/how-your-wristband-knows-you-are-stressed/</guid>
<pubDate>Mon, 09 Nov 2026 08:00:00 +0000</pubDate>
<description>Your wristband can’t read your mind: it listens to the tiny gaps between heartbeats. Here is how HRV becomes a stress figure, and what it can’t know.</description>
<content:encoded><![CDATA[<h2 id="a-heart-that-is-not-a-metronome">A heart that is not a metronome</h2>
<p>Put two fingers on your wrist and count. Your heart seems to beat with the regularity of a clock, but it doesn't. If we measured the exact time between one beat and the next, we would see it changing constantly: 850 milliseconds, then 910, then 870… Those differences, just a few hundredths of a second, have a name: heart rate variability, or HRV. And they are the main clue a wristband uses to tell whether you are under strain.</p>
<p>The curious thing is that a relaxed heart is <em>more</em> irregular, not less. When you breathe in, your pulse speeds up a little; when you breathe out, it slows down. This swing, called respiratory sinus arrhythmia, is normal and a sign of a well-tuned system.</p>
<h2 id="the-accelerator-and-the-brake">The accelerator and the brake</h2>
<p>Who pulls those strings? The autonomic nervous system, the part that runs on autopilot. It has two branches that work like a car's accelerator and brake.</p>
<p>The sympathetic system is the accelerator: it gets the body ready to act in the face of a threat, an exam or an urgent email. It raises the pulse and makes it more even, like a drum keeping time. The parasympathetic system, which travels mainly along the vagus nerve, is the brake: it takes over when we rest, digest or sleep, and lets the rhythm fluctuate freely.</p>
<p>When we keep our foot on the accelerator for a long time, variability drops: that is the footprint of stress that can be read from the outside.</p>
<h2 id="how-the-light-on-your-wrist-picks-it-up">How the light on your wrist picks it up</h2>
<p>The wristband has no chest electrodes: it shines small LEDs onto the skin and measures how much light comes back. With every heartbeat, the vessels fill with a little more blood and absorb more light. This draws a wave, and the algorithm finds the peak of each pulse to work out the time between one and the next.</p>
<p>From those intervals it calculates, for example, RMSSD, which compares each interval with the next one and sums up how much they change on average. A high value usually means the brake is on; a low one, that the accelerator is in charge.</p>
<h2 id="the-big-enemy-moving">The big enemy: moving</h2>
<p>Here is the catch. If you walk, climb stairs or simply gesture while you talk, your pulse rises from pure physical effort and variability falls even if you are perfectly calm.</p>
<p>That is why the wristband also has an accelerometer, the same kind of sensor that counts steps. When it detects that you are moving, it discards or corrects those stretches. The stress figure makes most sense in still moments: sitting at the computer, on the sofa or in a meeting.</p>
<h2 id="your-number-is-yours-alone">Your number is yours alone</h2>
<p>There is no "correct" HRV number for everyone. It changes with age, genetics, fitness and even the time of day.</p>
<p>That is why algorithms don't compare you with an ideal, but with yourself. During the first days, the wristband learns what your heart looks like at rest, especially at night. From then on, the stress figure shows how far you are from your own normal. It is a relative scale, more like a personal thermometer than an exam grade.</p>
<h2 id="what-the-wristband-cannot-know">What the wristband cannot know</h2>
<p>One thing, though: the heart does not tell emotions apart. Nerves before speaking in public and the thrill of watching your team score can leave a very similar footprint. The wristband measures how activated your body is, not whether you feel good or bad.</p>
<p>Things we don't usually link to stress also play a part: an afternoon coffee, a heavy dinner, a glass of wine, a cold coming on or a bad night. They all press the accelerator a little. So when the figure is high for no obvious reason, the explanation is sometimes on your plate or on your pillow.</p>
<h2 id="read-the-curve-not-the-moment">Read the curve, not the moment</h2>
<p>A single reading says little. What matters is the pattern: whether your figures always rise on Monday mornings, or whether a walk after lunch brings your pulse back to calm.</p>
<p>Want to try it? Sit down and breathe slowly for two minutes: in for four seconds, out for six. You are pressing the brake on purpose, and the curve often shows it within minutes.</p>]]></content:encoded>
<category>stress</category><category>HRV</category><category>sensors</category><category>how it works</category>
<enclosure url="https://duckdik.com/blog/img/2026-11/estres-hrv-pulsera-2a5a4bdf.webp" type="image/webp" length="0"/>
</item>
<item>
<title>Clock change 2026: we slept 26 minutes less that night</title>
<link>https://duckdik.com/blog/en/clock-change-2026-we-slept-26-minutes-less/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/clock-change-2026-we-slept-26-minutes-less/</guid>
<pubDate>Thu, 05 Nov 2026 08:00:00 +0000</pubDate>
<description>On the changeover night we slept 26 min less than a normal Saturday. Data from 21,941 nights in five countries show how the body adapted.</description>
<content:encoded><![CDATA[<h2 id="an-extra-hour-and-26-fewer-minutes-of-sleep">An extra hour… and 26 fewer minutes of sleep</h2>
<p>On Sunday 25 October we put the clocks back and, in theory, gained 60 minutes of sleep. What did we actually do with that gift? A DuckDik analysis of <strong>21,941 nights</strong> from <strong>1,569 people</strong> in Spain, Italy, France, Germany and Portugal, between 10 October and 1 November 2026, gives a counter-intuitive answer: on the night of the change we slept <strong>26 minutes less</strong> than on a normal Saturday.</p>
<p>It is not a clock error. It is what happens when the body and the calendar are not in sync.</p>
<h2 id="the-changeover-night-without-the-60-minute-bonus">The changeover night, without the 60-minute bonus</h2>
<p>In the early hours of Saturday 24 to Sunday 25, people wearing the <a href="/shop/duckdik-sense/">DuckDik Sense wristband</a> went to bed at roughly the usual time: around <strong>00:34</strong>. Had they used the extra hour, they would have slept more. Instead they woke earlier than you would expect, and the average result was <strong>7 hours 27 minutes</strong> of sleep, versus <strong>7 hours 53 minutes</strong> on a typical Saturday.</p>
<p>The comparison used <strong>777 people</strong> with data that night and on previous Saturdays, each measured against their own habits. Rest was also a little more fragmented:</p>
<ul>
<li>Nights under seven hours of sleep rose from <strong>27% to 39%</strong>.</li>
<li>Night awakenings rose from <strong>1.5 to 1.7</strong> per night.</li>
<li>Time spent awake during the night rose from <strong>18 to 22 minutes</strong>.</li>
</ul>
<p>In short: we were given an hour and, on average, we did not spend it sleeping.</p>
<h2 id="monday-almost-as-if-nothing-had-happened">Monday, almost as if nothing had happened</h2>
<p>What stands out is how fast “official” sleep recovers. On <strong>Monday 26</strong> average duration was <strong>7 hours 13 minutes</strong>, only <strong>four minutes</strong> below usual. People went to bed <strong>five minutes earlier</strong> and got up <strong>eight minutes earlier</strong>, on the new clock.</p>
<p>In other words: we adjusted our routines to the clock in a day. The body, meanwhile, was still catching up.</p>
<h2 id="the-working-week-did-not-lose-accumulated-sleep">The working week did not lose accumulated sleep</h2>
<p>Over the following days, many users kept going to bed and waking a little earlier than usual. The gap closed gradually. But one key figure stands out: across the <strong>working week</strong> average duration was <strong>7 hours 17 minutes</strong>, <strong>the same</strong> as in the reference period. There was no accumulated sleep debt. The change mainly moved schedules, not total rest.</p>
<h2 id="spain-adapts-sooner-germany-takes-longer">Spain adapts sooner; Germany takes longer</h2>
<p>The overall trend was similar, but the return to normal was not the same everywhere:</p>
<ul>
<li>In <strong>Spain</strong>, the usual wake-up time returned in about <strong>three nights</strong>, though people still went to bed about ten minutes earlier for part of the week.</li>
<li><strong>Italy</strong> needed around <strong>four nights</strong>.</li>
<li><strong>Germany</strong> showed the longest adjustment: waking about <strong>twelve minutes earlier</strong> all week and sleeping, on average, <strong>six minutes less</strong>.</li>
<li>In <strong>France</strong> there were barely any schedule changes.</li>
<li>In <strong>Portugal</strong>, people tended to go to bed earlier and sleep about <strong>nine minutes more</strong>.</li>
</ul>
<h2 id="early-birds-and-night-owls-do-not-react-the-same-way">Early birds and night owls do not react the same way</h2>
<p><strong>Early birds</strong> were almost back to normal by Monday, with gaps of only one or two minutes from their usual pattern. <strong>Night owls</strong>, by contrast, carried a small lag all week: they went to bed about <strong>eight minutes earlier</strong> and woke about <strong>six minutes earlier</strong>. Both groups, though, lost a similar amount of sleep on the changeover night.</p>
<p>“The clock change gave us an extra hour, but we did not use it to sleep. The most surprising thing is that, although we reset the clock in one night, our body needs several days to adapt. And almost all of this happens without us noticing,” says Alberto Lorente, co-founder and CEO of DuckDik.</p>
<h2 id="how-it-was-measured">How it was measured</h2>
<p>The study is observational: 21,941 nights from 1,569 DuckDik Sense users in ES, IT, FR, DE and PT, from 10 October to 1 November 2026. Each person was compared with their own prior habits on equivalent weeknights. Incomplete or inconsistent records were excluded, and the midnight clock shift was corrected. The results describe this sample, not necessarily the whole European population.</p>
<p>The full press release, with the country table and methodology, is in the <a href="https://duckdik.com/en/prensa/notas/cambio-de-hora-resultados-2026/">DuckDik press room</a>.</p>
<p><em>Alberto Lorente</em></p>]]></content:encoded>
<category>sleep</category><category>clock change</category><category>data</category>
<enclosure url="https://duckdik.com/assets/prensa/cambio-hora-2026/cambio-hora-2026-es.png" type="image/png" length="0"/>
</item>
<item>
<title>Why a screenless wristband lasts 45 days on one charge</title>
<link>https://duckdik.com/blog/en/screenless-wristband-battery-45-days/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/screenless-wristband-battery-45-days/</guid>
<pubDate>Sat, 31 Oct 2026 08:00:00 +0000</pubDate>
<description>Why a smartwatch needs charging daily but a screenless band lasts 45 days: what drains it, how Bluetooth saves power and why the phone does the maths.</description>
<content:encoded><![CDATA[<h2 id="a-gadget-that-spends-most-of-the-day-asleep">A gadget that spends most of the day asleep</h2>
<p>A smartwatch usually needs plugging in every day or two. A screenless wristband such as the DuckDik Sense can last up to 45 days on a single charge. The trick isn't a magic battery but a very simple idea: a device that spends most of its time asleep and only wakes up just long enough to measure and store what it has measured.</p>
<p>To understand it, think of the battery as a small piggy bank. What matters is not so much its size as how many coins slip out every minute. In a wearable there are four big energy drains: the screen, the processor (the “brain” that does the sums), the radio that talks to the phone and the light sensor that measures your pulse. A long-lasting band closes off the first and tightens the other three.</p>
<h2 id="the-screen-the-big-guzzler">The screen, the big guzzler</h2>
<p>In 2017, a team of researchers followed 27 people who wore a smartwatch in their everyday lives for 106 days and measured where the energy went. The screen took around 30% of the total, almost as much as the processor. And that's even though the watch was only fully “awake”, with the screen at full brightness, 2% of the time.</p>
<p>Why does something so small use so much? Because a screen is thousands of points of light that have to be switched on and redrawn many times a second, and because every time you check the time, the processor has to wake up to draw the image. Without a screen, that cost disappears and the processor has far fewer reasons to wake up.</p>
<h2 id="a-light-that-flickers-instead-of-staying-on">A light that flickers instead of staying on</h2>
<p>The sensor that measures your pulse works with light: a few LEDs light up the skin and a tiny electronic eye measures how much light comes back (we explain it in <a href="/blog/en/pulse-with-light-ppg/">Pulse with light</a>). The curious thing is that this light isn't on all the time. It flickers in very brief flashes, many times a second, and the sensor only “looks” during those flashes.</p>
<p>What's more, many bands adjust how often they measure to what you're doing: on the sofa there's no need to measure as intensely as in the middle of a run.</p>
<h2 id="bluetooth-that-talks-now-and-then">Bluetooth that talks now and then</h2>
<p>The band isn't “on the phone” with your mobile all day. It uses Bluetooth Low Energy, a system designed precisely so that the two devices meet at set moments, swap a little data and go back to sleep. According to the standard, these meetings can repeat anywhere from every 7.5 milliseconds to every 4 seconds, and if there's nothing new to report, the band can skip some of them.</p>
<p>Between meetings, a typical Bluetooth chip can drop to a consumption of around one microamp, a millionth of an amp. It's like two friends who, instead of chatting all day, meet for a second every now and then to catch up. That's why these bands usually store what they measure in their memory and send it in batches.</p>
<h2 id="the-wrist-collects-the-phone-thinks">The wrist collects, the phone thinks</h2>
<p>The other big key is sharing out the work. Interpreting sleep stages, stress or recovery takes a lot of calculation. Instead of doing it on the wrist, with a tiny battery, it's done on the phone, which has a battery many times larger and gets charged every day anyway. Raw signals stay on the phone; the account syncs already-summarised figures to servers in the EU.</p>
<h2 id="why-charging-less-also-looks-after-the-battery">Why charging less also looks after the battery</h2>
<p>There's a curious side effect. Lithium batteries age partly with every charge cycle. If you charge a device every day, that adds up to about 365 charges a year; with 45 days of battery life, only around 8. Fewer cycles usually means a battery that loses capacity more slowly, although time and heat also play a part.</p>
<p>Real battery life depends on how the band is used. Measuring blood oxygen all night, recording long workouts with continuous heart rate or getting lots of notification vibrations empties the piggy bank sooner. And intense cold temporarily reduces the energy a lithium battery can deliver, so on a winter hike it's normal to see the percentage drop a little faster.</p>]]></content:encoded>
<category>battery</category><category>sensors</category><category>how it works</category>
<enclosure url="https://duckdik.com/blog/img/2026-10/pulsera-sin-pantalla-bateria-45-dias-d340757a.webp" type="image/webp" length="0"/>
</item>
<item>
<title>How your band guesses whether you&#x27;re in deep sleep or REM</title>
<link>https://duckdik.com/blog/en/how-your-band-guesses-deep-sleep-or-rem/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/how-your-band-guesses-deep-sleep-or-rem/</guid>
<pubDate>Mon, 26 Oct 2026 08:00:00 +0000</pubDate>
<description>Without seeing your brain, a band slices the night into 30-second chunks and guesses your sleep stage from pulse, HRV and movement. Here&#x27;s how it works.</description>
<content:encoded><![CDATA[<h2 id="while-you-sleep-your-wrist-takes-notes">While you sleep, your wrist takes notes</h2>
<p>Every morning the app tells you how much deep sleep and how much REM you got. Curious, because the band can't see your brain or watch your eyes. So how does it know? The answer mixes indirect clues, statistics and a healthy dose of humility.</p>
<h2 id="how-a-sleep-lab-does-it">How a sleep lab does it</h2>
<p>To classify sleep "for real" there is polysomnography. The person sleeps with sensors that record brain waves (the electroencephalogram, or EEG), eye movements and muscle tension. Afterwards, a technician reviews the night in 30-second chunks and gives each one a label: wake, N1, N2, N3 or REM, following the rules of the American Academy of Sleep Medicine (AASM).</p>
<p>N1 and N2 make up light sleep. N3 is deep sleep, with slow, wide brain waves. In REM the brain becomes very active, the eyes move quickly and the muscles go almost completely slack.</p>
<p>Fun fact: not even the experts always agree. In an AASM programme where more than 2,500 technicians scored the same recordings, average agreement was 82.6%, dropping to 67% for deep sleep.</p>
<h2 id="the-clues-a-band-reads">The clues a band reads</h2>
<p>Without EEG, the band plays detective with what it can pick up from the wrist:</p>
<ul>
<li><strong>Movement.</strong> The accelerometer detects whether you move. It's the best clue for telling sleep from wake, but it says little about the stage.</li>
<li><strong>Pulse.</strong> An optical sensor follows your heartbeat. In deep sleep the pulse drops and becomes regular; in REM it gets restless.</li>
<li><strong>Heart rate variability (HRV).</strong> These are the tiny differences in time between one beat and the next. They reflect the tug of war between the nervous system's "brake" and "accelerator".</li>
<li><strong>Breathing.</strong> Breathing leaves a mark on the pulse: when you inhale the heart speeds up a little and when you exhale it slows down. From that, the breathing rhythm can be inferred: calmer in deep sleep, more irregular in REM.</li>
<li><strong>The time.</strong> REM is more common in the second half of the night.</li>
</ul>
<h2 id="slicing-the-night-into-30-seconds">Slicing the night into 30 seconds</h2>
<p>With those signals, the algorithm imitates the technician: it splits the night into chunks of about 30 seconds (eight hours gives you 960). For each chunk it calculates a handful of numbers, and a model trained on nights recorded at the same time with polysomnography decides which stage is most likely: awake, light, deep or REM.</p>
<p>That's why bands talk about "light" sleep instead of N1 and N2: without brain waves, telling those two apart is almost impossible.</p>
<h2 id="how-close-is-it-to-the-lab">How close is it to the lab?</h2>
<p>It's an estimate, not a polysomnography. The studies say so:</p>
<ul>
<li><strong>Telling asleep from awake works well.</strong> In a 2021 study of seven consumer devices, all of them recognised at least 93% of the sleep chunks, but only between 18% and 54% of the time spent awake.</li>
<li><strong>Stages are harder.</strong> With 106 people and 440 nights, a model that used movement alone got the stage right in 57% of chunks; adding pulse, variability and time of night raised it to 79%.</li>
<li><strong>Devices differ a lot.</strong> A 2023 study of 11 devices found agreement with the lab ranging from "moderate" to barely "slight".</li>
<li><strong>Restless nights are the hardest.</strong> With fragmented sleep, they all make more mistakes.</li>
</ul>
<p>In perspective: if human technicians agree around 83% of the time, an algorithm that can't see the brain reaching 79% is remarkable. But every single night has its margin of error.</p>
<h2 id="what-to-do-with-the-number">What to do with the number</h2>
<p>What matters is the trend. If one day shows 40 minutes of deep sleep and the next an hour and ten, it could be your night or the algorithm. But if over several weeks you see your deep sleep drop whenever you have a late dinner, that is useful information.</p>
<ul>
<li><strong>Compare weekly or monthly averages</strong>, not single days.</li>
<li><strong>Look first at the most reliable data:</strong> what time you fall asleep, how long you sleep and how regular you are.</li>
<li><strong>Use the stages as a clue</strong> to ask yourself questions, not as a verdict.</li>
<li><strong>If you sleep badly over a long period</strong>, the conversation is with a healthcare professional, not with the app.</li>
</ul>
<p>In a screenless band like the <a href="/shop/duckdik-sense/">DuckDik Sense</a>, which also estimates sleep stages, the interesting part comes when you look at several weeks in a row.</p>]]></content:encoded>
<category>sleep</category><category>sleep stages</category><category>sensors</category>
<enclosure url="https://duckdik.com/blog/img/2026-10/fases-sueno-pulsera-54d4eb82.webp" type="image/webp" length="0"/>
</item>
<item>
<title>We didn&#x27;t discover mathematics. We are mathematics</title>
<link>https://duckdik.com/blog/en/we-didnt-discover-mathematics-we-are-mathematics/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/we-didnt-discover-mathematics-we-are-mathematics/</guid>
<pubDate>Mon, 19 Oct 2026 08:00:00 +0000</pubDate>
<description>Not even its creators fully understand how AI reasons. That raises a bigger question: what if we ourselves are mathematics discovering itself?</description>
<content:encoded><![CDATA[<h2 id="what-even-their-creators-dont-understand">What even their creators don't understand</h2>
<p>There's something about artificial intelligence that I find especially fascinating. It isn't simply that it can answer questions, write texts or solve problems. It's that it is developing capabilities its own creators hadn't fully anticipated.</p>
<p>Yesterday I listened to an interview on Lenny's Podcast with Emmanuel Ameisen, a researcher at Anthropic who specialises in understanding what happens inside large artificial intelligence models.</p>
<p>The surprising thing is that their own creators don't fully understand how they work inside.</p>
<p>We know how to build and train them, but we don't always understand how they reach certain conclusions or why they develop capabilities that nobody has explicitly taught them.</p>
<h2 id="systems-we-grow-not-programs">Systems we grow, not programs</h2>
<p>Ameisen explains that these models are more like systems we grow through learning than traditional programs in which we have defined every behaviour.</p>
<p>Researchers are discovering that the models build internal representations of concepts, establish relationships and develop strategies that nobody has programmed directly.</p>
<p>All of this happens through mathematical operations. Not because the models invent new mathematical laws, but because training organises billions of parameters until it produces mechanisms that we are still trying to understand.</p>
<p>It's an extraordinary situation: we have created these systems, and now we need to investigate them to discover how they have learned to reason.</p>
<h2 id="a-question-that-goes-beyond-ai">A question that goes beyond AI</h2>
<p>And while I was listening to the interview, a question came to me that goes well beyond artificial intelligence.</p>
<p>What if mathematics weren't simply a tool we discovered to understand the world? What if we ourselves were mathematics?</p>
<p>What if we've got it backwards?</p>
<p>We have always thought that humans discovered mathematics to explain the universe. But the relationships it describes existed long before us.</p>
<p>The planets followed their orbits before Newton. Atoms interacted before chemistry existed. And our brains work through physical processes that also obey those same laws.</p>
<h2 id="a-language-a-face-a-brain">A language, a face, a brain</h2>
<p>Think of something as everyday as learning a language. Our brain discovers patterns, establishes relationships and ends up building sentences it has never heard. Curiously, artificial intelligence models do something similar through completely different mechanisms, based on mathematical operations.</p>
<p>Or think of our ability to recognise one face among thousands. We do it almost instantly, without being able to explain how. Today an artificial neural network can develop a similar ability by learning relationships across enormous amounts of data.</p>
<p>That doesn't mean our brains work the same way as an artificial intelligence. But it does raise something fascinating: perhaps intelligence depends less on the material we are made of than on the relationships established between its components.</p>
<p>And if those relationships are mathematical, what are we really?</p>
<h2 id="a-hypothesis-not-a-certainty">A hypothesis, not a certainty</h2>
<p>Perhaps mathematics isn't simply a language we invented to describe reality, but part of the structure of reality itself.</p>
<p>This hasn't been proven. That something can be described mathematically doesn't necessarily mean it is mathematics. But what we are observing with artificial intelligence makes it worth considering.</p>
<p>Because there is something extraordinary about a set of mathematical operations being able to learn, reason and develop capabilities that not even its creators fully understand.</p>
<h2 id="mathematics-discovering-itself">Mathematics discovering itself</h2>
<p>What if we are another example of that same phenomenon, one that emerged through millions of years of biological evolution instead of artificial training?</p>
<p>Perhaps we are not beings who discovered mathematics from the outside.</p>
<p>Perhaps we are a part of the universe that, organised in a particular way, has acquired the ability to understand the rules that made it possible.</p>
<p>And perhaps we haven't discovered mathematics. Perhaps we are mathematics discovering itself.</p>]]></content:encoded>
<category>artificial intelligence</category><category>mathematics</category><category>brain</category>
<enclosure url="https://duckdik.com/blog/img/2026-10/no-descubrimos-las-matematicas-778d10b0.webp" type="image/webp" length="0"/>
</item>
<item>
<title>What time Spain falls asleep: data from 36,700 real nights</title>
<link>https://duckdik.com/blog/en/what-time-spain-falls-asleep-data/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/what-time-spain-falls-asleep-data/</guid>
<pubDate>Wed, 14 Oct 2026 08:00:00 +0000</pubDate>
<description>Spain falls asleep at 00:30, an hour later than Germany. Data from 36,700 nights in five countries, ahead of the 25 October clock change.</description>
<content:encoded><![CDATA[<p><em>Updated 14 October: wake-up times and weekend data corrected.</em></p>
<h2 id="spain-falls-asleep-at-0030-the-latest-of-five-countries">Spain falls asleep at 00:30, the latest of five countries</h2>
<p>On Sunday 25 October the clocks go back one hour and the yearly question returns: does Spain live on the wrong schedule? Data from 36,727 nights recorded by wearables between July and October 2026 give a concrete answer. In Spain the average time of falling asleep is 00:30, around one hour later than in Germany. Wake-up time, on the other hand, is delayed somewhat less, by about 50 minutes. The result: four out of ten nights in Spain last less than seven hours.</p>
<p>The data come from 1,322 users of the DuckDik Sense wristband in Spain, Portugal, Italy, France and Germany. Of them, 454 live in Spain and account for 12,523 nights.</p>
<table>
<thead>
<tr>
<th>Country</th>
<th>Falls asleep</th>
<th>Wakes up</th>
<th>Actual sleep</th>
<th>Nights under 7 h</th>
<th>Steps per day</th>
</tr>
</thead>
<tbody>
<tr>
<td>Spain</td>
<td>00:30</td>
<td>08:05</td>
<td>7 h 20</td>
<td>41 %</td>
<td>10,307</td>
</tr>
<tr>
<td>Portugal</td>
<td>00:22</td>
<td>08:01</td>
<td>7 h 21</td>
<td>39 %</td>
<td>9,123</td>
</tr>
<tr>
<td>Italy</td>
<td>00:04</td>
<td>07:38</td>
<td>7 h 19</td>
<td>40 %</td>
<td>9,758</td>
</tr>
<tr>
<td>France</td>
<td>23:57</td>
<td>07:49</td>
<td>7 h 35</td>
<td>33 %</td>
<td>8,367</td>
</tr>
<tr>
<td>Germany</td>
<td>23:25</td>
<td>07:14</td>
<td>7 h 29</td>
<td>35 %</td>
<td>8,595</td>
</tr>
</tbody>
</table>
<p>Mind the maths: bedtime and wake-up time mark the time spent in bed, while actual sleep leaves out the minutes it takes to drop off and the awakenings during the night, which is why it comes out a bit lower.</p>
<h2 id="why-the-time-you-fall-asleep-matters-not-just-how-many-hours">Why the time you fall asleep matters, not just how many hours</h2>
<p>The difference between countries lies less in how much people sleep than in when. A Spanish user sleeps about ten minutes less than a German one, a small difference. But they do it with their sleep shifted one hour towards the small hours.</p>
<p>That has a practical consequence. Wake-up time is set by work and school, and it shifts less than bedtime. Whoever falls asleep late has less margin: any night that runs a little long ends up too short. That is why in Spain 17 % of nights do not even reach six hours.</p>
<h2 id="the-weekend-another-time-zone">The weekend, another time zone</h2>
<p>From Sunday to Thursday, the average time of falling asleep in Spain is between 00:13 and 00:21. On Friday night it moves to around 00:50, and on Saturday to one in the morning.</p>
<p>Four in ten users shift their sleep by one hour or more between weekdays and the weekend. Specialists call it "social jet lag": on Monday morning the body is still on Saturday's schedule, just as after a flight across time zones.</p>
<h3 id="young-people-fall-asleep-latest">Young people fall asleep latest</h3>
<p>Between 18 and 29, the average time of falling asleep is 00:52. From 60 onwards, 23:31. People over 60 also sleep the most (7 h 39), although they wake up more often during the night.</p>
<h3 id="women-sleep-about-20-minutes-more">Women sleep about 20 minutes more</h3>
<p>Women sleep 7 h 40 on average and men 7 h 20. Women also fall asleep about half an hour earlier.</p>
<h2 id="spain-walks-more-and-naps-less-than-people-think">Spain walks more and naps less than people think</h2>
<p>Two figures that break stereotypes:</p>
<ul>
<li><strong>Spain is the country that walks the most</strong> of the five: 10,307 steps a day on average, 20 % more than Germany (8,595) and 23 % more than France (8,367).</li>
<li><strong>The daily siesta is a myth.</strong> Only 11 % of days include a nap. In Germany and Italy it is 9 %.</li>
</ul>
<h2 id="what-will-happen-with-the-clock-change-on-25-october">What will happen with the clock change on 25 October</h2>
<p>Turning the clock back one hour means, in theory, one more hour of sleep that night. In practice, the body does not change schedule in a day: for several days it tends to fall asleep and wake up at the usual time.</p>
<p>"We have spent years debating the clock change with opinions. Now we have data from thousands of real nights to see what really happens," explains Alberto Lorente, co-founder and CEO of DuckDik. "We Spaniards do not sleep much less than our neighbours, but we go to sleep later than anyone. In the week of the clock change we will measure, night by night, how long the body takes to adapt, and we will publish the result."</p>
<p>The second analysis, with the real effect of the clock change by country, will be published in the first week of November.</p>
<h2 id="how-the-data-were-obtained">How the data were obtained</h2>
<ul>
<li>Aggregated, anonymous data from users of the DuckDik Health app with at least seven nights recorded between 1 July and 12 October 2026.</li>
<li>Times are local in each country. The country is the one where the user spends the most time.</li>
<li>Every user counts the same, whether they have more or fewer nights recorded.</li>
<li>It is not a representative sample of the population: these are people who have bought a wellness wristband, and 72 % of those who state their sex are men.</li>
<li>Sleep is estimated by the wristband from movement and pulse.</li>
<li>The calculation was repeated only with users who sync daily, without manually corrected nights and without open customer-service tickets. The conclusions are the same, with differences of a few minutes.</li>
</ul>
<p>The full press release, with charts and the country table to download, is in the <a href="https://duckdik.com/en/prensa/notas/sueno-cambio-de-hora-2026/">DuckDik press room</a>.</p>]]></content:encoded>
<category>sleep</category><category>trends</category><category>clock change</category>
<enclosure url="https://duckdik.com/blog/img/2026-10/portada-sueno-2026-v4-en-4cb7feb1.webp" type="image/webp" length="0"/>
</item>
<item>
<title>Pulse with light: how a wristband reads your heart from outside</title>
<link>https://duckdik.com/blog/en/pulse-with-light-ppg/</link>
<guid isPermaLink="true">https://duckdik.com/blog/en/pulse-with-light-ppg/</guid>
<pubDate>Tue, 13 Oct 2026 08:00:00 +0000</pubDate>
<description>How a wristband knows your pulse using only light: what photoplethysmography (PPG) is, how it works out pulse, HRV and SpO2, and why it sometimes errs.</description>
<content:encoded><![CDATA[<h2 id="a-tiny-torch-and-an-eye-that-counts-reflections">A tiny torch and an eye that counts reflections</h2>
<p>A wristband knows your pulse using only light: it shines a small green light into your skin, a sensor measures how much light bounces back and, because blood absorbs light, every heartbeat changes that amount very slightly. By counting those rises and falls per minute, the device works out your heart rate. The technique is called photoplethysmography, although almost everyone knows it by its initials: PPG.</p>
<p>Turn an activity tracker over and you will see a few little lights (LEDs) and one or two dark dots: these are photodiodes, a kind of electronic eye that turns light into an electrical signal. No needles, no electrodes: just light going into the skin and light coming back out.</p>
<h2 id="why-can-light-see-blood">Why can light "see" blood?</h2>
<p>When light hits your wrist, it meets skin, fat, bone and blood vessels. Almost all of that is constant: your bone does not change size from one second to the next. Blood, however, moves. With every heartbeat a small wave arrives that slightly swells arteries and capillaries. Because haemoglobin (the protein that gives blood its red colour) absorbs green light very well, more blood means less light gets back to the sensor.</p>
<p>The result is a huge, almost fixed signal with a tiny ripple on top that rises and falls with your heart. The hard part is not switching on the light, but rescuing that ripple.</p>
<h2 id="from-a-wave-to-a-number">From a wave to a number</h2>
<p>The sensor measures light many times per second and draws a curve of small hills, one per heartbeat. The software does three things: it <strong>cleans</strong> the signal (like removing the hum from a recording), <strong>finds the rhythm</strong> (if the wave repeats 1.2 times per second, that is 72 beats per minute) and <strong>checks</strong> that the result is reasonable.</p>
<p>Measuring the exact time between beats also gives you <strong>heart rate variability (HRV)</strong>: those intervals vary by a few milliseconds, and that variation reflects the balance of the nervous system. It needs a very clean signal, which is why it is usually measured at night or at rest.</p>
<h2 id="why-it-sometimes-fails">Why it sometimes fails</h2>
<p>The blood ripple is so small that almost anything can hide it:</p>
<ul>
<li><strong>Movement.</strong> When you run, the band shifts and creates false waves. That is why it carries an accelerometer: knowing how your arm moves, it can subtract that pattern.</li>
<li><strong>Outside light.</strong> Wearing it snug (not tight) and a couple of fingers above the wrist bone helps a lot.</li>
<li><strong>Cold</strong>, which narrows skin vessels, and <strong>tattoos</strong>, because ink also absorbs light.</li>
</ul>
<p>Rule of thumb: data is more reliable at rest and asleep than in the middle of a sprint.</p>
<h2 id="red-and-infrared-oxygen">Red and infrared: oxygen</h2>
<p>With two more colours you can estimate <strong>SpO2</strong>, the oxygen in your blood. Haemoglobin with and without oxygen absorbs light differently, so pulse oximeters compare red light (about 660 nanometres) and infrared light (about 940 nanometres) and translate the result into a percentage. On the wrist this is harder than on a finger, so wellness wristbands show it as an overnight trend, not as a clinical test.</p>
<h2 id="ai-from-numbers-to-sentences">AI: from numbers to sentences</h2>
<p>Using pulse, HRV and movement, models trained on thousands of nights <strong>estimate</strong> sleep stages (in deep sleep the heart is slower and more regular; in REM, more variable) or your stress level. They are estimates, always compared with your own baseline. Artificial intelligence helps above all to <strong>interpret</strong>: it turns hundreds of points into clear sentences such as "you slept less deeply than your average today".</p>
<p>In some wristbands, raw signals stay on the phone and only summarised figures sync to servers in the EU; we explain it on <a href="/wearable-datos-europa-rgpd/">our page about data and GDPR</a>. To dig deeper, the <a href="https://en.wikipedia.org/wiki/Photoplethysmogram" rel="noopener">Wikipedia entry on photoplethysmography</a> is a good start, and for questions about your wristband there is the <a href="/support/">support centre</a>.</p>]]></content:encoded>
<category>pulse</category><category>sensors</category><category>SpO2</category>
<enclosure url="https://duckdik.com/blog/img/pulso-con-luz-ppg/pulso-con-luz-ppg-2195b335.webp" type="image/webp" length="0"/>
</item>
</channel>
</rss>
