Ok, so I’m using resting voltage as a proxy for charge retention.
Measured over 4 months on 7 of 8 cells (4S2P) pulled from a now dead laptop’s battery pack. The cells are 13 years old.
Chart is interactive, with actual voltage and date values shown by mousing over the data points.
I was curious and maybe others are too. Enjoy.
ive got some ~12 year old laptop pulls laying around…probably need to check their voltages 
Funny, I have a old Sony 18650 (pulled from a Viao laptop we bought new in 2003) that I was using today to test Lexel’s rear FET switch I was building up.
Pulled it out of the fridge stash where I stored it maybe 5 months back or so with about a 3.7V charge at 33 degrees. After running about 5 min or so on turbo direct driven into three xpl hi leds on my hunting C8Fs (old 18650 hosts), I was putting it all away. But I thought I’d check voltage and it showed 3.69V!
Not too shabby for a 17+ year old 18650 eh?
May or may not be related… I notice that old high-resistance batteries will have voltages bounce back to higher voltage quickly.
In fact, for some very old cells, I notice that after draining them (eg. do a “discharge” without charging back), the voltage will quickly increase, some to even 3.8-3.9v. I notice good batteries (especially the good high-drain ones), when the voltage goes down, the voltage will stay at the low voltage.
So, what I’m guessing is that your old Sony 18650, could have voltage dropped very low while in use, however, it quickly bounced back to the higher voltage right after pulling it from the flashlight.
Hum, I might have to check the IR on the old Sony for kicks and grins 
Side note: I pulled six from the pack about 7 years ago, and I have always gifted them with lower power lights. But I remember testing a few I had left a few years back and they made it to about 2000mAhs which I thought was amazing.
Cool chart phouton.
I have some 10+ year old batteries that I took out of a gently used laptop. They haven’t been charged in over a year and are resting at 4.15 volts. They have a 1,900 mAh capacity and only use 1 of them in my low powered S2+ UV flashlight. Not sure why I keep the other 5.
Good point about the resting voltages not necessarily corresponding to actual capacity. It’s unclear what can be inferred by the voltage alone. It is a convenient measurement though, especially over time, as opposed to a capacity test. I don’t have equipment to check IR though that would be good to know and track over the lifetime of a cell.
I’m open to suggestions for other measurements that can be taken with a multimeter, or other ways of presenting the information. I also made another plot showing the drop in voltage per day over time. I didn’t post it here since it doesn’t see too interesting.
It’s just so satisfying to make full use of resources, even if they’re not at peak performance anymore. But I’m a bit loathed to use these 7 cells (the 8th is being used) since it would disrupt my measurements
Updated the voltage measurements on the laptop pull cells, now ongoing for nearly 6 years. They are all still over 3.7V from a single initial charge. Also, these cells are now old enough to vote.
The Google Sheet link in the first post is interactive and automatically updates with new data, but it is not embedding in the discourse forum post. So, here is a static image:
I can post a chart if it seems useful, but for now i will just relate overall observations on my laptop pull cells. I have very many, but really three larger groups which have valid characteristics from when I first started measuring them. The date that I first characterized them, unless otherwise noted, is August 2019.
The first set is 16 Samsung icr18650-28a cells. I have two date codes, One date code is July 2015 and the other is June 2013. Eight cells from each date code. All 16 cells are at a voltage of around 4.13 volts. So those cells are at or near puberty.
The second set of cells are a set of eight LG cells LGABC2165, their date code is September 2013. All of those cells are sitting right at 4 volts, with the lowest at 3.98 V.
Finally a set of 6 LG cells from October 2011, LGDAS31865. The voltage on these cells are all around 4.07 volts. One cell is higher ar 4.12V, and the one that had the lowest capacity when first tested in 2019 is at 4.07V.
If this forum and I are both still around in another 10 years, i can report back on these cells when they’re getting close to as mature as the ones that you’ve tested.
I do have some Sony cells with date codes as old as 2006 and 2008, but I got them in May of 2024, so not long enough to really see how they’ll do long term. So far the voltage is over 4.1V on all 14 of those cells, which are in pairs still.