Of course it's poor inventory management - OP was willing to go back and buy the $259 S fare, but UA panicked and opened L, thereby losing out on $100 :)

I'm being a little facetious of course, but it is illustrative of exactly the dilemma UA has to face when managing their prices. Their only motivation here is profit; they want to charge you as much as you're willing to pay to fly, while still being able to sell all of their seats possibly to other pax who won't pay as much. A perception of consistent pricing just doesn't enter into the equation. Why would they care about that? That being said, actual fare tables typically move slowly, and so if you regularly fly a particular route you can become quite familiar with the prices. For example, over the last four months I've flown the exact same C fare between SFO and WAS something like 7 times.

It's worse with gasoline prices sometimes changing [increasing] 15-20 cents in 8 hours. And then all the stations in a 20 mile radius follow suit.

Etiene, Aliquot - it is poor management because UA's algorithms told them they could garner $259 one way... then 16 hours later that algorithm realized it was not getting the sales required, and they had to drop prices 3 levels to get some. That is poor inventory management. Similar to what findark says. It was good revenue management by me, although that was nothing but sheer luck. My point is - and I realize this is solely in one case - if the airlines have to move down 3 buckets in such a short period of time to garner sales then they priced the fares inappropriately at the start. If they were selling out seats at $259, then they would not drop it 3 levels 16 hours later.

Simple analogy which demonstrates this has nothing to do with United or its accounting/inventory/pricing management.
You have a rental car/SUV on empty and returning to airport and have the following options:
a. You know there is a gas station at the corner of the rental car row which charges $4.00 p/gallon but it is easy/simple and no hassle. With your full size vehicle you are looking at $60 for 15 gallons
b. You are 20 miles from airport and can exit the interstate and buy gas at your current location which is $2.50 p/gallon. But, not sure if it will remain full by the time you get to the rental car row and whether traffic/tieup enroute forcing you to stop again or charged by rental agency for returning less than FULL (which may exceed the savings as they will charge for a gallon plus fueling service fee if you are under Full). Cost at station is 37.50 + risk ????.
c. You don't want to choose A but want something closer so you choose to drive in circles or down the service road seeking a gas station closer to airport but not the surcharged gouging price. This puts you about 10 miles (5 miles away and back) from the roadway to the airport down a long empty road. There may to may not be a gas station at the end of the road (you did see a sign but no clue where the station). No clue of the price per gallon when you get there. Cost on this ????, extra time minimum 12-15 minutes (10 miles drive), so if your time is worth $25 p/hr you have spent $7.50 of time.

In the above, think of A as someone who simply wants to buy a ticket to the destination and doesn't really want to check back or price match as they take the trip once a year and that is it. B is the person monitoring prices taking the risk it may go up and stay but they prefer to buy it low. And C is the person in option paralysis - they will rather risk time and stress over what will be minimum vs. maximum pricing models.
Airlines know the above people exist and there is always a % of people who fall into A, a % who fall into B and C is the unknown. Make money on A (which is the premium pricing on Friday), fill up the empty seats with B and C depends how much is left over.

Looks like customer preferences change hourly too.

Sure, but the bigger problem here is that while a human analyst who was dedicated to watching this flight in specific at all points in time might have been able to make a slightly better call and get your $100, UA has literally tens if not hundreds of thousands of flights on the schedule at any given time and it's not cost effective to monitor each one closely. Almost all of these inventory decisions are made by a computer, and the software is not even all that sophisticated (they rolled out a new management system recently and were proud of it on their quarterly call). Getting "perfect" inventory on a given route takes an investment, so there is a cost/benefit tradeoff.

Since there is no program that can yet predict human behavior, we are sadly doomed to "suffer the slings and arrows of outrageous fortune" with our ticket prices.:)

It's quite possible (and maybe likely) that United didn't change anything at all and that some other customers did exactly what OP did -- booked and cancelled within 24hrs (or let a farelock expire). If they held the last of the $159 bucket tickets (thus only showing $259 when you originally booked), and cancelled them, the $159 price would show again.

If you have the email receipts, what was the fare class of the original and later ticket?

That isn't how inventory is managed. There is no guarantee that cancelling a ticket will reopen inventory in a lower fare bucket. Sometimes it does; sometimes it doesn't.

The more important point is that there are a limited number of seats at $159. At that point, there are not repricing all seats. It's just that the lowest price of one or a few seats is $159.

That's why it's called "dynamic"
 

The conditions in place when you made your first rez were different when you came back to check. Revenue management is not about any individual pax, but about the airline's ability to maximize revenue. Chances are the algorithm takes into account situations such as you describe (some pax in 24h window may cancel higher fare and go for lower) in determining how to price. If you are exceptionally price sensitive book whenever then keep taking serial advantage of the 24h window - but make sure to snag the new, lower fare prior to cancelling the previous. It is not unheard of that between your searching for an itinerary and pulling the trigger on it the fare expires. Not common, but it happens. .

I suspect most of us have been unpleasantly surprised by a price rise resulting from an expired fare, missing an AP requirement, or zeroing of an inventory bucket. Probably more than once.

You'll never outsmart the UA algorithms over the long run as you don't have the data UA has - you got lucky this time.

You should always look a gift horse in the mouth as you can tell a lot about the animal's health that way. Someone may be trying to "gift" a sick horse.

Of course they are; that is what businesses do.

Weren't you playing games too, shopping for the lowest fare, canceling and rebooking, trying to "milk" UA by having them provide a service to you for the lowest possible price and having them leave money on the table? (Remember you were in the end willing to pay the higher fare.)