← The Science Bowl Lab

Canonical Science Bowl runs on six categories — biology, chemistry, physics, math, earth & space, and energy. (NSBA swaps in computer science, a non-canonical substitution; everything canonical below is measured on the six, and CS is flagged separately wherever it appears.) The six look interchangeable from the stands — points are points — but they are nowhere near equal. They come up at wildly different rates, they're wildly different to staff, and, most importantly for how you play, they demand two fundamentally different kinds of answer. This page is the long version of what the buzz data says about each.

01Which categories matter

volume, independent leverage, and how hard each is to cover

Start with the trap. If you just correlate a team's strength in a category with whether it wins, every category comes back at roughly +0.7. That looks like a profound finding and is actually a nothing: good teams are good everywhere, so strength-in-anything correlates with strength-overall, which correlates with winning. Raw category correlation measures team quality, not category importance.

The honest question is independent leverage: after you control for a team's overall scoring, does being extra strong in category X still predict extra wins? That isolates whether the category is a lever you can pull to win beyond your general level, rather than just another readout of it. Pair that with two facts you can't ignore — how many points the category is worth (its share of tossups) and how hard it is to even field a capable answerer — and the six sort into a clear order.

Read down that table and the ranking writes itself, with a different reason behind each slot:

That last finding is the counterintuitive one, and it's worth sitting with. The instinct is to draft a "math ringer" early because math feels like a specialist skill. The data says the opposite: because math strength is almost entirely a proxy for the general talent you're already measuring, paying up for it specifically is redundant. You get the math for free by drafting good players; spending a premium pick to over-index on it is negative-leverage.

02Recall vs computation

the two kinds of question, and how differently you have to play them

Here's the idea that reorganizes everything above. The six categories don't just differ in importance — they split into two mechanically different types of question. Biology, chemistry, earth & space, and energy are recall: you either know the fact or you don't, and if you know it you can buzz the instant you recognize it. Math and physics are computational: you have to hear the setup, then do work, then answer. That intuition — bio is "know it or you don't," math is "figure it out and execute" — is correct, but it shows up in the data in a more specific and more useful way than you'd guess.

The first thing to notice is tempo. Computational questions get buzzed roughly 0.14 of the question later — you cannot answer a physics problem until you've heard the whole setup, so there's no recognize-and-jump. Math is the extreme: it's buzzed at 0.93 of the way through, later than any other subject, because you have to hear every term of the problem before you can even begin to compute. Recall, by contrast, gets buzzed early and often — the moment your brain matches the clue, your hand moves.

The second thing is format. Computational questions are almost always short-answer (75% vs 53%), and math specifically is only 15% multiple-choice. That closes two doors at once: you can't recognize-and-jump on an early clue, and you can't guess from a list of options. In the recall categories you frequently get "which of the following…" — and there, on a bad night, a coin-flip guess after the options are read is a live play.

Now the twist that flips the naive intuition. You'd expect the "hard, do-work" categories to be riskier — lower conversion, more disasters. The opposite is true. Once a strong player commits to a math or physics buzz, they convert as reliably as recall (63.1% vs 62.0%) and the question goes dead far less (4.5% vs 8.3%). Mastery makes computation deterministic. If you can do the problem, you do the problem — there's no fuzzy half-memory, no "I think it's the one with nitrogen." The hurdle in math/physics isn't accuracy-once-committed; it's the tempo-and-format tax of having to hear the whole thing and produce an exact answer.

So the strategic upshot cleaves along the recall/computation line. In the multiple-choice-heavy recall categories — Energy is 56% MC, Earth & Space 49% — guessing after "which of the following" is positive expected value, so aggression pays; fire on your best guess. In math and physics, do the opposite: hold, read to the very end, and buzz only once you've actually executed the computation. Early aggression there just buys you negs on problems you hadn't finished solving.

03The bonus by category

the +10 payout, and where it's easiest to cash

Winning a tossup earns the right to a bonus — a +10 question, no penalty, no bounceback to the opponent. Across the field bonuses are converted roughly 59–65% of the time, which makes them the majority of all scoring. But conversion is far from uniform across the six, and the spread is enormous: bonus conversion swings 27 points from top category to bottom.

Chemistry (72%) and Biology (66%) sit at the top — deep, learnable subjects where a strong answerer who won the tossup is very likely to also have the bonus. Earth & Space (45%) sits at the bottom, with Physics (52%) and Math (58%) in between. That Earth & Space is simultaneously the highest-volume category and the worst-converting on both the tossup and the bonus is exactly why it is the single biggest "you either know it or you don't" swing in the game: it's the most points on the table and the least reliable to actually collect.

Zoom out and the bonus is where games are decided. Roughly two-thirds of all points scored are bonus points — the tossup is the contest, the bonus is the prize — and team bonus conversion correlates +0.63 with winning. A team that wins the buzzer race but leaves bonus points on the floor is beating itself; a team that cashes its bonuses turns a modest tossup edge into a comfortable margin.

04The graveyard: dead by category

where points go unscored — and why the answer is field-dependent

A dead tossup — nobody on either team answers correctly — is a category's points vanishing into the void. Each dead tossup forfeits roughly +9 to +10 points (the tossup plus the bonus it would have unlocked), so a category's dead rate is a direct measure of how many points it quietly erases from the game.

Earth & Space is the graveyard — 23.6% of its tossups die unanswered, comfortably the worst, with Biology (19.4%) a clear second. Physics, Math, and Chemistry cluster in the mid-16s, and Energy (14.9%) is the most reliably answered. This lines up with everything above: the categories that are hardest to convert and worst on the bonus are the same ones that most often go dead. Earth & Space's breadth-and-obscurity problem shows up here too — it's simply the category where the marginal question most often stumps everyone.

The practical reading: don't treat any category's difficulty as a universal constant. In a broad, open field, Earth & Space is where points die and covering it well is a real edge. In a concentrated elite pool that's already strong there, the same category becomes safe and the scarcity edge moves elsewhere. Scout the field you're actually in.

05Caveats