Not to cause a ruckus, but I mix old and new batteries. I have a 20-battery bank of 6V Trojans. The black ones cost about 5500 or 6000 usually. They are stamped with a date mark on the negative terminal - letters A-L corresponding to months 1-12 followed by the last digit of the year. This is a "new" date, but sometimes you see batteries in the store with the NEXT month's stamp; so it is not the manufacture date. There is also a date stamped on the BOTTOM of the battery case, they say, but in 20 years of buying 6V Trojan batteries, I have never once looked(!) if you want to see that, get a mirror! I won't buy batteries more than a month old - but that is getting harder to find as more of the country gets 24 hour power and inverter use slowly declines.
So: the old batteries have some sulfation on the plates. That means they have lowered storage capacity. If there is a new battery in the bank, it cannot charge to full capacity as easily because the batteries around it reach their "full" voltage with less stored wattage - the charging current is lower due to those full batteries, and the newer battery charges very slowly (but it is still charging). When the batteries discharge, the older batteries drop voltage earlier and cause the new battery to have to do more work and discharge deeper. The next time it recharges it has the same problem again with slower charging once the others are full.
The first time this "change all batteries together" policy collided with reality was about a year after I installed the bank. One battery lost one cell. I measure the specific gravity of each cell each 60 days and top off water. Back then, I was also monitoring the daily recharge pretty closely, and the day I lost the cell, I saw the bank stop reaching full charge for no obvious reason other than a battery problem. Was this a 6000 peso problem or a 120,000 peso problem? I decided I could live with a little suffering on the part of the one new replacement battery if it kept the other 19 operational.
Once you go down the road of mixing batteries, it is not easy to see a way back to "purity". From time to time, I replace a battery when a cell goes bad. I alway try to revive the battery first with a few zaps of 12V and boiling (theory being that a bad cell arises from material bridging between positive and negative plates and shorting the cell - vigorous gassing in the cell may dislodge the bridge material). If this works, the three cells of the zapped battery approach the same specific gravity within a couple of days - and that has happened a couple of times. One such "saved" battery has continued to serve normally for four years since the zapping. But usually it doesn't help. So I have some newish batteries and some oldish batteries. If I wanted to become pure I would get rid of the oldish batteries with little pain, but the newish ones would laugh at me as they departed - (see ya', sucker!).
Well, my 20 batteries are keeping the lights on: We use an average of 6kWh daily from a 2400W solar array. Most of the power usage is spread out over 24 hours - not concentrated in daylight hours: refrigeration, freezing, and water recirculation in a gray-water system. We don't change usage when it is cloudy - which is not an uncommon situation here in the mountains. If the power goes low after three days of cloud cover, we have a 5kW propane fired generator to make up the difference. That generator runs about 30 hours per year on average.
