These are also strange results. Although read speeds are back up somewhat, they're still slower than the single drive. Write performance increases further over that of the two-drive array, but I can't fathom why the read performance is inferior to that of just one drive. Again, there should be another performance increase from that of two drives in all areas. The story continues with a four-drive RAID 0 array:

Now, an ideal four-drive RAID 0 array should have four times the performance of a single drive, and I don't think it's unreasonable to expect that, even in practice, a four-drive array should have over 250% of a single drive's performance. But these numbers fall far short of those expectations. Some of these numbers are better than those turned in by the three-drive array, but some are worse. SiSoft's Sandra favors this array compared to a single drive, but the other benchmarking programs produce varying results in such a comparison. Again, I have no good explanation for these confusing numbers.