Some of the terminals that look like aluminum are aluminum, some tinned copper alloy. To tell the difference, scratch it. Aluminum is light grey all the way through. Tinned alloy is usually yellowish underneath. Tinned also accepts solder well.
A good solder joint will not crack under vibration or temperature extremes a typical car will see. A wire solder joint is not meant to hold the wires together mechanically. The wires are to be twisted or wrapped or tied or whatever for the mechanical joint. The solder is to bind the electrical aspect of the wires. Circuit board soldering can mechanically hold light components to the electrical connection, but larger components are separately soldered via a ground plane with a mechanically twisted tab or crimped or glued.
A "cold" solder joint will go bad usually cracking. A "cold" solder joint occurs when the joint is moved before the solder cools to solid temperature. A poor mechanical twist of wire will allow movement for a "cold" to possibly form. A good solid mechanical twist will not. A dirty solder joint will fail as well.
Pipe soldering is a bit different. The solder joint is an extremely thin gap between the pipe and fitting. They do provide a mechanical joint, but only when the gap is small. There is also a lot more surface area than a wire joint. Pipe joints usually also fail due to a cold or dirty joint.
I'll trust a good solder joint more than a good crimp.
Oh, also there was some bad solder alloys produced during the transition from lead based to lead-free solder, that was the cause of alot failures as well.
Most of the stuff that I quoted above regarding solder joint failures WAS prompted by the military. Even the highest quality level solder joints, namely military and aerospace (not just aircraft, but things like satellites and space capsules and space shuttles) have had solder joint failures, in all types of printed circuit boards, and not just with heavy components. It continues to be a leading failure mode to this day, even in solder joints that are made under carefully controlled conditions, with as close to perfect soldering processes, on solder joints that are X-Ray'd, in aerospace and in the military. This is NOT a new occurrence, it has been happening for a long, long time, pretty much since we have been using solder as a way of electrically joining components and wires. Studies on solder joint integrity have been going on for at least 80 years, if not longer. Solder joint failures even occur, to this day, on tiny surface mount resistors that weigh just a gram or two.
A lot of mechanically secure, properly made solder joints also fail. What happens is that the solder joint fractures. Then moisture and dirt get into the crack. If what caused the initial fracture continues (and in the majority of the time, it is vibration), the fracture will continue to expand, moisture and dirt will continue to get into the crack, and a failure will occur over time. How long it takes to fail is a function of both the amount and intensity of the vibration, and to a lesser extent, temperature swings. In all cases, the cause of the failure is the same, it is the inflexibility of the solder itself that is the issue. Solder is NOT flexible; it cannot absorb vibration, nor can it expand and contract at the same rate as the components that is is joining together. Note: this even disregards improperly made solder joints, whether they are cold, dirty, just bad, or poor quality solder (i.e. where the proportion of the tin/lead is off, or even in the newer non-lead solder).
Solder joints WILL crack in automotive conditions. The automotive environment is not particularly benign; there is vibration (engine and other internal components, plus whatever the road/tire interface contributes) and cars are typically used in all sorts of weather (-40 to +45C outside air temperature, and typically, -40 to about +100C or so under-hood). This is a lot more difficult than say, a PC, or an LCD TV, which may cycle a few times during transportation, and spend most of their lives inside a house at about 22C, no vibration at all. Components often times have to be specified for automotive use, and cost more. Even so, one of the leading causes of electrical failures in consumer electronics, automotive electronics, industiral electronics (central office switches, cellular base stations, power plant control panels), and even military/aerospace electronics is solder joint failures.
The ideal crimping tool is the ratchet crimper, which is what is called a controlled cycle tool in that it is not left to the opinion of the operator whether the crimp was sufficiently compressed or not. You keep squeezing the handles until the tool releases, and you have a good crimp. Takes the person out of the equation in terms of crimp quality, so that you don't get great crimps to start with, but as time goes on, the crimp quality suffers due to the operator getting tired, or bored. You will also note that the ratchet crimper makes a flat crimp, whereas the non-ratchet crimpers make a smaller indent that is often times curved, or more of a groove. This makes it a lot less iffy as to whether the crimp was made evenly, so if the wire is slightly misaligned, it isn't an issue with a ratchet crimper.
As with all things, don't cheap out on the crimp terminals. The amount of engineering that has gone into the crimp terminal would amaze most people. I highly doubt that the dollar store terminals, or the ones on Allie or Bang Good have had that kind of effort put into them.