OK, that's probably my fault and I didn't mean it that way. Let's all be nice.
I've fiddled with RAID on and off, and I've come two the conclusion that there are two reasons to build a RAID and that any other reason is probably a mistake.
Reason 1) To play with it and learn something, or have fun.
Reason 2) To solve a specific business need. RAID used to address disk speed and capacity issues, but nowadays there are better ways to achieve either. RAID still does an excellent job of ensuring that your data remains available after a single (or even double) disk drive failure, but you are still vulnerable to motherboard / controller failure.
-----------
RAID 0 is, by definition, not RAID at all. RAID once stood for "Redundant Array of Inexpensive Disks;" now it stands for "Redundant Array of Independent Disks." Really. All RAID levels were supposed to increase fault tolerance. RAID 0 reduces fault tolerance and is not Redundant in any way.
RAID 0, or striping, will allow you to roll the capacity and speed of two disks into one volume. It will have twice the capacity of a single drive and roughly 1.5 to 1.9 times the speed. However, if anything goes wrong with either drive, the odds are that you will never see any of your data again without sinking thousands of dollars into it. RAID0 is a good place to put game levels if you can't afford an SSD that large, but have backups of everything and be prepared to accept falling back to the backups.
By all means build RAID 0 and try it out. It will be faster than a single drive but slower than an SSD. Just be aware of the increased risk that you are accepting. Also, you may have to tweak your OS before setting up the RAID set, or the OS won't boot after you change the motherboard controllers to the mode that supports RAID.
All the higher (real) RAID levels combine N physical drives into one volume that has the capacity of N-x times the capacity of one of the drives. N=2, x=1 is RAID 1, mirroring, where each drive contains byte-for byte the same data. If one drive fails, your system goes chugging along happily. While both drives are up, some controllers will split reads across both drives, improving your read speeds.
Higher levels of RAID require the calculation of parity information, which either loads your CPU or requires a specialized RAID controller that does the parity calcs. For example, a RAID 5 with 4 drives will write one block to each of 3 drives and the parity block to the fourth. If any drive fails, all three blocks can be reconstructed from what remains - either you have the three raw blocks, or two plus the parity and you can recalculate the missing block.
-------------------------
In my mind the greatest weakness of RAID is that it is not standardized at the byte and configuration level. This means that a RAID set of drives built on one controller may not be readable on a different brand, or even model, of controller. If your controller dies, and it was old, you may never see your data again (although there is information at the top of the storage forum on working around this).