I don't think any of these options are necessarily the best for approaching "Quiz Bowl" science. Problem with Quiz Bowl science is that it is almost always asking conceptual or definition questions and this only works in biology out of the major sciences. Both Chemistry and Physics are primarily equation-driven. Now, what's wrong with that? You'll be spending more time in a lecture learning how to apply problem solving techniques than actually learning simple concepts, formulas, and quantities. For example, you could learn kinetic energy inelastic/elastic collisions and this would take a decent amount of time by watching a lecture and genuinely understanding the concept. I could memorize that kinetic energy is simply denoted as T in the Hamiltonian and buzz in on that. It's simply not worth your time to watch lectures.
To get to what you should actually do, I would start with the Stanford Culture Page's science links at http://ai.stanford.edu/~csewell/culture/
. From here, I would go on to Protobowl, make a room and just look at each HS Science question that comes up. Learn the content in that clue and especially pay attention to laws and equations. Certain ones will come up all the time mid-way through (Biot-Savart Law, Sackur-Tetrode Equation, etc). For smaller categories, like math, you can get a really good chunk of questions by simply going to quizbowldb and typing in mathematician's name and reading over all their clues.
So I took a look at that link, and it's really not very good. Some examples:
London forces - weak attractive forces in molecules; vary as 1/d7; only intermolecular forces among
Dipole-dipole interactions - attraction of opposite partial charges; vary as 1/d4
Hydrogen bonding - H to F, O, or N; like dipoles
These are really bad descriptions of intermolecular forces. First off, the radial dependences are misleading--people 99% of the time will talk about the energy of the interaction, not the force, and the energy associated with an LDF goes as 1/r^6 (also, seeing as studying this way will cause you to NOT have AP physics and AP calculus knowledge, you won't know the relationship between forces and potentials and will just be confused). Additionally, the dipole radial dependence is wrong (dipole-dipole interactions go off as 1/r^6, not 1/r^4) Furthermore, hydrogen bonds are not "like dipoles," they are a special type of dipole-dipole interaction.
Most importantly though, it doesn't actually explain what any of these things are! What is a dipole? (furthermore, what is an instantaneous dipole or an induced dipole?) What causes a London dispersion force? What is polarizability, and how does it relate to LDFs?
Enthalpy - heat content
Transition state theory - activation energy to form transition state
I don't even think I need to elaborate on how terrible and woefully inadequate these "explanations" are.
I looked on the page for the explanation on orbitals--but there isn't one! How are you supposed to learn chemistry at all if you don't even know what an orbital is?
The Biology page
is just laughable. Reading these questions, sure, will allow you to answer those questions written on the page reliably. But when the same material is presented in a different format, or the clues are different, you're just going to get screwed! The physics page was comical--it's basically an equation sheet. I didn't look at any of the other pages, but I'm sure they're all equally awful as resources for actually learning things.
Your idea that you can get good at science by using Protobowl and memorizing old question content is flawed for many reasons. First off, if you don't actually understand the basic concepts behind chemistry/biology/physics (which can ONLY be acquired through reading a textbook, taking a class, preferably of AP level, or looking at online teaching resources), you're not going to understand the things you see in tossups. So, when another tossup presents the material in a different way, or uses slightly different clues, you're going to be confused and not earn the points. Sure, you can memorize that kinetic energy is sometimes symbolized T, but that's only going to get you points on terrible questions. [writers of housewrites, pay attention here: saying "this quantity is symbolized T in the definition of the Hamiltonian" is a terrible clue and you should never use it ever] Packet studying is useful for identifying the type of material which comes up, but inadequate for actually learning things about it. Furthermore, no matter how much packet studying you do, when someone comes up with a creative new clue that hasn't come up in packets before (or only rarely does so), you're screwed. At ACF Regionals yesterday, (admittedly not a high school tournament), I buzzed on a couple things that to my knowledge (and my computer's packet archive's knowledge, and the hsqb packet archive's knowledge) have never come up in quiz bowl before, based on knowledge I've acquired through classes. You just can't compete against good players with real knowledge on well-written questions.
As a conclusion, I'll state that I am not saying that you should never look at old packets ever or that packet studying is evil. Packet studying is a good way to augment your knowledge base once it is already in place. [and reading textbooks, taking classes, and looking at online teaching materials is the only way to create an adequate knowledge base] When you have real knowledge of something, it will make it easier for you to understand what is coming up in packets, and easier to remember things you learn in packets. Binary associations get lost when you don't remind yourself of them repeatedly, but conceptual understanding of something is difficult to forget, and will reward you more in the long run, both inside and outside of quizbowl contexts.