I think that the issue cryzymotor is encountering with cells that don't stay in balance is caused by a couple of things.
First, the cells being used probably don't have very good QC, so they don't match each other very well even within the same batch, because the defective ones weren't weeded out at the cell manufacturer.
Second, the cells are probably being used beyond their actual capabilities (regardless of what the vendor *says* they are capable of, which is often not realistic, and probably doesn't reflect what the manufacturer actually rates them at).
When quality cells are used, that have been sorted well by the manufacturer, with defective ones removed from the supply chain, and they are used within the actual capabilities of the cell (not the extremes they might be able to do momentarily, but what they can realistically sustain without changing the characteristics of the cell, heating them up, etc), then during their expected lifetime they may not *need* any balancing.
As one example, I am running EIG NMC 20Ah cells, rated at 5C, and peak usage is around that, while typical usage is much lower (probably around 1C or less). There are balance wires on the pack but not attached to anything; no BMS, just occasional monitoring for testing. They stay balanced, even after years of use this way, using only bulk charging.
The only cells I've had any issues with were a single one that began to fail with higher internal resistance after several years, so it sagged more at higher currents, and three cells from a separate lighting pack that was accidentally run down to 0v once, but which still worked at lower currents; it's even been run down very low once after taht and only one of those cells died (puffed up) because it actually reversed voltage a bit. The other two cells still work at their reduced current rating, and are still in use now (with the original high-Ri cell in the same pack, also working at that reduced rating).
THey also do not heat up on their own, from use, in any of the testing I've done, because they are actually capable of more current delivery than the manufacturer rates them for, unlike many of the "generic" cells of various types which are rated higher than they can really deliver just for marketing reasons, as those manufacturers probably have no reputation to protect, I'd guess, and no warranties to worry about honoring, unlike EIG (and others that realistically spec their cells).
As an example of improper specifications, I've got a pack of 13s4p 18650 cells, which while probably genuine, are spec'd by the pack vendor for much higher currents than they can realistically sustain in that pack configuration (nowhere to shed their heat, for starters), and since the BMS allows this current, they'll get out of balance when used at the higher rates pretty easily, and the BMS will detect LVC on a cell and shut off the pack, even though it actually has plenty of capacity left if it were used at the lower rates the cells are realistically capable of, instead. The cells also wouldn't get hot if they were rated realistically.
Then there's various "who knows" stuff like the RC LiPo packs I've got, where even when new they didn't stay balanced, even when used at much lower rates than the pack vendors spec them for (and manufacturer specs are unknown, as the actual cell manufacturers are not known), and various cells could die (puff up) even just sitting there at median charge levels, while not in use.
So there are a number of factors in play