there is a feature request to address this here: https://forum.inaturalist.org/t/option-to-hide-other-observations-on-maps-observation-detail-page-identify/2246.
I’ve been trying to locate some screenshots of old observation page maps to see if this has always been a problem, but no luck. so i’m not sure if what you’re describing is new or not…
i’ll get back to that, but first, let me describe generally how a map is structured, as i understand it. when you look at a map, you see a rectangle with a basemap and some markers on top. the basemap is delivered not as one image but as a bunch of smaller tiled images. as you pan the map, additional tiles are retrieved to fill in the places that come into view. (this tile approach prevents the system from having to either deliver one giant map image or constantly update the image to fit what’s shown in the map rectangle.)
you might think that the markers (ex. pins) are rendered as individual objects on the map, but they actually are delivered as image tiles, too. you can seemingly interact with individual markers because there’s another invisible UTFGrid tile layer tied to the marker layer that defines spots where some action (like clicking) can be taken. (the tile approach is used to display the markers because on some maps, there could be thousands or even millions of markers, and it could be very inefficient to try to send over and render on the client machine thousands of markers. it’s more reliable to send over tiled images of the markers instead.)
finally on the observation page (and i think also on the Identify page, and a few other places), the map there also has a larger marker that represents the specific observation referenced by the page. that marker is rendered as an individual object, not as part of a set of tiles (because there’s only one central observation on that map).
so right now, the big individual marker is basically replicating a marker that is part of an image tileset, and that’s why you see the “extra” marker. but think of what it would take to remove the “extra” marker. the system would basically have to dynamically generate a unique tileset that excludes a particular observation. technically that’s possible, but it is probably super inefficient.
for example, suppose you look at two different observations A and B that are near each other and are both identified to Dicot level. when you look at their respective observation pages, the maps right now can use the same Dicot marker tileset for both observations. however, if you exclude the “extra” marker, you’d have to generate 2 unique tilesets – one that is Dicots excl. A and another that is Dicots excl. B. not efficient.
i mentioned earlier that i was trying to figure out how the maps used to look. i suspect that in the past, the big markers were simply shifted down a little bit to cover the smaller corresponding tileset marker. that might have looked okay back in the day when the markers at higher zoom levels were relatively large circles. but now, the markers at higher zoom levels are smaller grid squares, and it might look odd if the large marker tip was landing below the area covered by the corresponding grid square.
now, let’s get to the reason you think the extra marker is annoying:
i don’t think this is necessarily the right logic. even if there’s just one pin there, that one pin could still represent multiple observations, if the multiple observations have (roughly) the same location. (in such a situation, the system appears to give you the details of the latest observation at that location if you click on it.) if you zoom in, sometimes you’ll notice that a single pin will be divided into multiple pins.
also, because the pin is so big, it could be hiding one or more small pins behind it even if you zoom in a lot. for example, look at this observation: https://www.inaturalist.org/observations/45431429. if you click around, you’ll see there are multiple pins hiding behind the big pin. so you would still need to zoom in and out or click around to reveal those hidden pins anyway, and sometimes those hidden pins still don’t reveal that there are multiple observations there.
at the end of the day, i’m thinking this is all a minor annoyance, and maybe a better understanding of how things are structured allows you to look past it. ether way, i doubt this would be considered a bug.