There are maybe two different segments with the same name, one counted 4 times and the other three times. In case your activity is visible to everyone could you post the URL for it for further checking?

​@Jan_Mantau - This is an odd one.  Here is a link to the segment in question: Parliament Stair Climb | Strava Run Segment in (Old) Ottawa, Canada.  I couldn’t pull up his specific activity since his profile is private, but I did pull up a few others in the leaderboard.  It looks like the results are actually for the same segment but it is registering them multiple times in some cases.  For example, in some it will register the climb (which is the actual segment), but then will also register another attempt when they go back down and then immediately back up again but still count the second up as a separate attempt.  That is likely how he ended up with 7 attempts when he only went up 4 times.  It counted 4 ups, but also counted the times when he went down-up as well.  It probably has something to do with it having a hard time knowing if they actually completed the segment since the start and end points are close to the start and stop of the set of stairs.  It is also a set of stairs with tight switchbacks, so any GPS drift and it can think you are further along the course.  

Here is another person’s activity that shows this happening over and over again: Stair season is back 🥵 | Run | Strava.  I am guessing it happens more when the GPS signal is poor.  The activities I see with a cleaner GPS track seem to only log the segment on the up like it should, but those with signal all over the place are getting picked up multiple times.

The activity from that other person is a nice example indeed. A segment with an incorrect GPS track, an activity with even worse GPS recording and Strava’s algorithm is overwhelmed. It’s really a bug then seeing that the same segment id is registered two times in the same timeframe.

I’ve seen a case when a short segment (a bridge across a river) was matched for someone going through that bridge in the opposite direction. The problem lies in the Strava segment matching algorithm that has very relaxed matching criteria. To match a segment both start and end have to be within something like 75 meters of the actual start and end, and a certain percentage of points have to match, but also within a similarly large radius. 

I’ve also seen cases when on a short uphill segment, someone turns around 2/3 way up, and their attempt still matches with a record time and they get a KOM.

In my opinion, considering that modern GPS devices are so much more accurate, segment matching threshold needs to tightened down. But I suspect the main reason Strava has it so much relaxed is because it is designed to match cyclists going at full speed and with Garmin Smart Recording being enabled, which may result in recorded points being fairly far apart - as far as 50-70 meters. That’s probably why the threshold is so large. But then the same threshold applies to running and we get quirky results like the one above.

Here is the link to my activity: 

https://strava.app.link/tnsaq4iBtUb

The other commenters already clarified the situation pretty well, though.