I don't understand how you see a contradiction in them at all - actually I don't really know what you mean.
Image complexity and encoding parameters are seperate things. I was referring to the fact that image complexity can have a large impact on bitrate required to achieve a certain level of quality. So, for example, a "simple" 30 minute video may only require 100MB to look excellent, whereas a "complex" 30 minute video may still look bad if given 400MB.

Encoding parameters certainly aren't a perfect way to judge an encode (or even that great, to be honest), I just said it was (almost always) better than looking at a file size, and can be seen by clicking on the file link above. It's the job of the encoding application to determine the best way to encode a particular input video, taking into consideration the desired level of quality/filesize tradeoff or speed/compression tradeoff, device compatibility etc. Encoders like x264 also allow you to add some subjective biases (eg favour textures over lines, blurring/blocking tradeoff) so in theory you could tweak these for whatever content you're encoding, but I find that in practice, these can be highly subjective, the defaults are usually quite sensible and you can't stray too far away from them, meaning that their impact is limited.

A note about rate control. Most people are probably familiar with the bitrate rate control, where you supply a specific target bitrate and the encoder will try its best to guess quality levels to meet this target. Bitrate based encodes allow you to predict the final file size of the file, and is useful if quality is a secondary consideration to, for example, fitting a certain number of episodes on a CD/DVD.
When file size isn't a concern though, you're probably better off not specifying a bitrate and using a rate factor rate control (crf in x264 terms) instead. With this, you specify a level of quality you're willing to lose (so a higher rate factor will give you lower quality). The encoder will then simply apply this value at the quality reduction (quantize) stage. Because of this, the file size can vary greatly and is usually difficult to predict, but you'll generally get a relatively consistent level of quality for the same rate factor.
You can get the rate factor used in the x264 parameters from the link above - in particular "rc=crf" (rate control = constant rate factor) and "crf=23.0" (rate factor value of 23 was used). Commie's version linked above also uses a rate factor, but crf=15, which means higher quality but larger file size.
Personally, a crf below 25 is quite acceptable for watching, as long as you're aren't trying to nitpick or look for defects. crf 15, I would consider to be transparent (indistinguisible from the source) to most people. YMMV
My own preference is to download the originals because I like the higher quality, so I certainly don't have a preference for these re-encodes. However I disagree with the statement that they are virtually unwatchable.