Skip to main content Skip to bottom nav

How is average wait calculated?

calmingOcean6858 October 29th, 2021

I'd be curious to see the algorithm behind how average wait is calculated. Does it mean "How long you can expect to wait on average between joining the back of the line now and getting matched with a Listener?" Or does it mean, "How long have the people currently in line been waiting, on average?" Or what?


The reason I ask is that I suspect the amount of time the average person spends in line is much longer than the listed average wait time.

5
vlostwithyouv October 29th, 2021

Mean median or mode?

QuietMagic October 29th, 2021

@calmingOcean6858

I'd thought it would be the 2nd one ("How long have the people currently in line been waiting, on average?"), but just from looking at the queue I can see that it can't possibly be that. The queue currently has an average wait time of 2:43 but half of the people in the queue have been waiting for at least 8 minutes.

I'm guessing that the calculation is based on taking the average wait time from accepted chat requests within a certain time range.

If that's the method, I think it's completely possible that this could be accurate considering that 1) sometimes chat requests get accepted instantaneously, 2) chat requests that have been in the queue for a long time probably cancel their request after waiting for a while.

3 replies
RarelyCharlie October 29th, 2021

@QuietMagic I agree. It must be a moving average that includes requests that are no longer visible in the queue. It's not very useful, really. The maximum wait time would be more interesting.

From time to time I've asked my computer to monitor the general request queue, and on seeing this thread I did it again today for an hour from 8:00 to 9:00 a.m. EDT.

In this hour, there were 233 requests, 224 of which left the queue. (At the end of the hour there were still 9 in the queue.) I have no way to tell whether someone left because a listener chatted to them, or because they just went away.

These 224 requests were made by only 115 members, of whom 94 were adults and 21 teens. 57 of the members appeared in the queue more than once.

The most requests were by an adult who joined 7 Cups today, and who entered and left the queue 10 times in the hour. (I thought that had been prevented.) This member's waiting times in the queue averaged just 12s.

The longest wait was by a teen member who joined 7 Cups years ago but hasn't been very active, who waited in the queue for 14½ minutes, then left for 7½ minutes, entered again, and waited 22:40. The nine longest waits were by teens.

Overall, for the 224 requests, the mean (ordinary average) time in the queue was 2:10. The modal (most common) time was only 4s (9 requests). The median was 36s—that is, half the requests were in the queue for 36s or less.

Maybe listeners often choose to chat to members who haven't been waiting long. Maybe members often join the queue by mistake and leave quickly, perhaps multiple times. It's difficult to interpret the numbers.

Charlie

2 replies
QuietMagic October 29th, 2021

@RarelyCharlie

Thanks for checking. 😊 That mirrors what I've seen sometimes when I open the GR queue. There will usually be a lot of requests that get added and then quickly disappear, and then there will be some other requests that have been in the queue for a really long time.

Assuming the general request response rate on the listener homepage (~95%) is accurate, that would mean nearly all of the chats that left the queue were accepted by a listener. If that's the case, I wouldn't know what exactly the thought process is behind listeners often picking people from the bottom of the queue instead of the top (e.g. maybe name or profile picture looks interesting, maybe they're picking based on topic, maybe people who have been in the queue for a long time have a higher chance of being ghost chats so they pick a more recent request, etc.)

AffyAvo October 30th, 2021

@RarelyCharlie Thanks for sharing that! I have suspected there was a change to the chat request limit, it seemed likely to me it was removed or increased.

load more
load more