Author: Michael George
Created: 26 Aug 2026
The adoption of COROS devices by GPS-Speedsurfing was based on a collaboration between the two parties.
COROS released firmware in Nov 2020 which introduced the Windsurfing and Speedsurfing activities. Whilst the Windsurfing activity is like many of the other fitness activities, Speedsurfing included metrics specifically for GPS-Speedsurfing, and recorded Doppler-derived speeds.
The APEX Pro and VERTIX were the latest models at the time, and they provided some really nice features to the community. Metrics such as 2 sec, 10 sec, 250 m, 500 m, and NM are great on the water. The phone app could also produce a nice session summary, near perfect for social media.
It was also possible to link a GPS-Speedsurfing account with a COROS account, so sessions would be automatically uploaded. It was a great user experience, and it simplified the process of posting sessions after the session, eliminating the need for a laptop and cables.
The early models (APEX Pro and VERTIX) used a GNSS receiver from Sony, which had mediocre GNSS performance at best. However, Locosys had stopped producing the GT-31, GW-52, and GW-60 so the community urgently required a successor.
The willingness of COROS to collaborate, and the convenience of their watches made them a sensible choice. They weren’t perfect, but the view was that they could be improved over time. It was only later when it become evident that the watches had a variety of issues.
The APEX 2 Pro, VERTIX 2, and VERTIX 2S initially had significantly improved GNSS performance, thanks to the Airoha AG3335M chipset. Sadly in May 2024, firmware issues started to occur and that is when things started to go downhill in terms of GNSS performance.
Some of the known issues are documented within this website; investigations, and firmware details. Some of the issues are not even documented, because frankly it takes a lot of time and effort, and I have limited spare time available.
One of the biggest issues in recent years has been that all of the watches have received firmware updates which ruined the speed data. Specifically, recording position-derived speeds (instead of Doppler-derived) causing unrealistic results, often by multiple knots, or more.
Another significant issue that remains unresolved is intermittent errors that produce “fantasy” alphas that can be wrong on GPS-Speedsurfing by up to 10 knots. This is thoroughly documented and it is clearly an issue with how COROS handle the GNSS data.
Lastly, lots minor bugs are present such as not understanding that gybes mark the end of runs. So, you can go and sail for an hour (or more) and if you are good at gybing the COROS watch (and app) will report a single run for the whole session.
There used to be regular discussion with the Head of Product Marketing and Support at COROS, but there has been no contact since April 2025. There has been no notice of new product releases (e.g. NOMAD or COROS APEX 4), or updates about the firmware issues.
The various firmware issues during 2024 continued to re-occur during 2025, and testing every firmware release is unrealistic. I have personally spent hundreds of hours working on COROS issues (probably 300+ hours), and it is very frustrating to see issues re-occurring.
Neither I, nor the guys at GPS-Speedsurfing have heard from our COROS representative for almost 18 months, and despite the early promise, it feels like the company has lost interest in our community, and incapable of resolving issues that we report.
Just recently there has been a discussion on Facebook, initially about the APEX 4 but various topics have subsequently been discussed up such as Doppler, and multi-band. That is what prompted me to write this page, so that some of the relevant information can be shared.
The COROS APEX 4 (42 mm and 46 mm) was released in Oct 2025, and the specification suggests that COROS may have switched to a different GNSS chipset. This adds uncertainty around the performance, and we have already seen that change is not always for the better.
My expectation is that the bugs present on the APEX Pro, VERTIX, APEX 2 Pro, VERTIX 2, and VERTIX 2S will still be present in the APEX 4. Establishing whether this is the case is rather time consuming though, and not an activity that I have undertaken as yet.
The green icon on GPS-Speedsurfing is often thought to represent a “Doppler” device, but that is not actually the case. It actually represents a device that has been approved for use on GP3S, considering the overall performance, presence of Doppler-derived speeds, and correct usage.
Satellites are orbiting the earth twice per day, whilst the earth is rotating, whilst we are windsurfing. The GNSS signals are affected by the Doppler effect, which essentially changes the frequency. In order to process the GNSS signals, receivers must determine the Doppler shift for each signal.
Without going into the gory details, knowledge about the satellites (e.g. position, orbit, velocity) and the observed Doppler shifts, it is possible to estimate the velocity of the receiver. The Doppler-derived speed is typically a lot more accurate than speeds derived from successive locations.
Multiband is not a pre-requisite for Doppler-derived speeds. It is also not a silver-bullet for speed accuracy, but the Airoha AG3335M chipset does support multi-band, and it demonstrates some great performance, assuming the COROS firmware is not broken.
In the context of watches, multi-band simply means they can use two signals from each satellite (e.g. L1 band, and L5 band). The signals on the L5 band are more powerful, and thus provide better penetration of tree cover, and some common building materials.
The L5 signals also have a higher “chipping rate” which results in greater precision when estimating the distance between the satellite and the receiver (aka pseudo-range measurement), somewhat akin to the original P(Y) code of GPS which was for military purposes.
Perhaps the most significant factor of using the L1 and L5 signals is to allow the ionospheric delays to be modelled, and largely eliminated from the navigation solution. The ionospheric delay is the most significant error source in the pseudo-range measurement.
There are other benefits such as some natural immunity to multipath error, and faster signal acquisition. There is no obvious benefit of multi-band for Doppler measurements, and the benefits are so slight that they are hard to quantify through experimentation.
A little known fact is that whilst a GNSS chipset may be capable of producing Doppler-derived speeds, its use is dependent on the activity type. For example, there is no benefit to a hiker knowing the speed of their watch oscillates between 0 and 10 km/h when hiking at 5 km/h.
Thus different activity types calculate speeds in different ways, and the vast majority of activity modes do not use the Doppler-derived speeds from the GNSS chipset. The Speedsurfing activity for COROS watches is the only activity that uses Doppler-derived speeds.
On a related note, Garmin watches also use Doppler-derived speeds for activities such as windsurfing, kiteboarding, waterskiing, etc. The quality of the Doppler-derived speeds is dependent on the GNSS chipset, filters, oscillator, and antenna design.
There are many layers in between the GNSS chip, and the results that you eventually see on website such as GPS-Speedsurfing. The activity mode has already been discussed, but the choice of file format is another thing where things can go awry.
COROS is the only watch manufacturer to include speed in GPX files, but it is not part of the GPX 1.1 standard, and thus often ignored by GPX imports. It is strongly advised that COROS users share + upload FIT exports, thus ensuring the Doppler-derived speeds are retained.
No other watch manufacturer includes speed in their GPX files, but FIT files always include the speed that is calculated for the activity. This goes back to the topic of using a suitable activity mode to ensure the Doppler-derived speeds are recorded, so Speedsurfing on COROS.
COROS and Garmin are the only two watch manufacturers known to provide activity modes suitable for our community. Suunto and Polar do not have any way to record the Doppler-derived data, and Apple is a another can of worms that I will not discuss right now.
Hopefully this has provided some relevant background information about COROS devices between 2020 and 2025. What happens about the APEX 4 is still to be decided. The GNSS performance is unknown, and reliability of future firmware updates cannot be predicted.
I have numerous GPS projects on the go for the wider community, so I haven’t investigated the APEX 4. I have plans to document how I assess GNSS performance of new devices, and the variety of issues that can be present. It’s not as simple as just comparing the top runs.
TBH, I don’t really know what to think of COROS at this time. The fact they have re-introduced the same issues over and over again does not fill me (or the guys at GPS-Speedsurfing) with confidence. We have also lost all contact with the company, which is not great.
I’ve not ruled out an investigation into the APEX 4, but I do have a lot of other projects on the go, which are summarised in the link above. The next couple of months are going to be super-busy for me, during which I won’t have any bandwidth for COROS investigations.
Like I said earlier, hopefully this provides some clarity and insights into why the APEX 4 has not been assessed as of yet. We all have busy lives, and many things competing for our spare time. If anyone owns a Motion (or ESP GPS) and APEX 4 then please reach out to me.
If you want to understand GPS / GNSS in more detail, I wrote a series of short articles on Medium.
There is also a more informal collection of technical articles on Github:
The above links are a tiny percentage of what I have investigated in recent years, perhaps 5 to 10% at most.