ThirdStats counts are always incorrect

[desrod]

I was hoping ThirdStats would show the details of how many mails per day/week/month/year, but the counts are always wrong.

Currently, ThirdStats claims I'm getting 9.2 emails per-day in the last 1 month.

But looking at Thunderbird directly, the counts are significantly higher than that:

I've dumped the cache, re-built, tried a new profile and so on, all with the same results.

How can I get ThirdStats to show accurate statistics?

[devmount]

Hi David,

thank you for this report. What do you mean by looking at Thunderbird directly? How do you create the view in your second screenshot? Also the numbers in your second screenshot seem odd. How can there be more emails in the last 7 days than in the last 14 days (which I assume must include the last 7 days).

ThirdStats only uses Thunderbirds official WebExtension APIs - so the counts reflect all messages retrieved over those APIs.

[desrod]

@devmount When I view the folder, threaded, Grouped by Sort, and sort by date, then collapse those folders, I get those counts. It's all built-in.

I received 118 emails today, not including some junk I deleted. When I constrain ThirdStats to look only at my Inbox, and only from 2023-06-04 to 2023-06-05, it says "Mails Total: 14", "Mails per day: 14.0", "Mails received: 13", and that's clearly false.

Where are these numbers coming from? How are these calculated?

[devmount]

First of all: I cannot reproduce your problem. I created the view like you did:

And compared with ThirdStats for this folder:

(sorry it's German) The numbers match on my end.

Where are these numbers coming from? How are these calculated?

Messages are retrieved via Thunderbirds official Webextension APIs and counted accordingly. All messages, that have a sender that is configured as identity in Thunderbird, are counted as sent, all others as received.

Things you can do:

1. Check that you have the correct account selected on the stats page
2. Check that you have the correct folder selected on the stats page
3. Check, if all messages of that folder are actually downloaded by Thunderbird (not only the message headers - I used to experience some number differences, when Thunderbird only had message headers stored due to performance reasons, which weren't delivered over webextension APIs)

[desrod]

I've confirmed this with 2 cleanly created profiles on 2 separate machines, using the same version of Thunderbird on both (macOS and my daily Linux machine), and the results were the same.

The counts were 10x to 100x below the actual count of mail in each of the selected folders. All mail was fully sync'd to the local machine, and it's the same whether I'm "offline" or online, when re-generating the statistics.

I have 8 different mail accounts that I use, and each of them has the same results, none of the counts were accurate across any of them. It's definitely consistent across OS/platform and accounts.

I think I've noticed at least one part of why this is thrown off. For my work account, I receive quite a bit less email on the weekends (~200 emails a day on weekdays, maybe 10-50 on the weekend days), and this throws the statistics off by quite a bit as I include prior years.

Having a "Days" count next to the Years | Months | Weeks would make this a lot easier to understand as well. I'll see if I can't PR that in. Excluding weekends from the parsing would also be useful, on the main stats screen (a simple checkbox: "Exclude weekend days" for example).

I think this might be why the stats are progressively off the further back the date constraint goes.

Thanks for testing for me!

[devmount]

This is interesting. Though I can't think of any way how this could be connected to weekends. Just to make sure: You don't have any warnings when processing like this?

Which email provider(s) do you use? Several users had problems with virtual folders from some providers (e.g. gmail) in the past.

[devmount]

Feel free to reopen this, if you have any news on this.

[desrod]

Thanks for your patience. The issue still persists with any account I have tested with (7 separate Gmail accounts).

Just looking at yesterday's incoming mail to my my most heavily-used account, there are 236 emails in my Inbox received on that date, where my email address is directly on the 'To:' line (not Bcc nor Cc, and not from a mailing list).

Searching exclusively in my Inbox, ThirdStats says I'm receiving 59.7 emails per-day (after clearing and rebuilding the cache). There are over 1,500 emails in my Inbox (intentionally not cleaning it out, in prep for a talk about email).

The counts just don't seem to be consistent.

Is there some debug mode I can run in and figure out precisely what math it's using to calculate these, and where it's going off trail and miscounting them?