Episode 12: The Full LR Attack Plan

Adrian Calloway

Here is the way most people lose three points on a Logical Reasoning section. They hit a nasty Parallel Reasoning question around number sixteen, they smell that they almost have it, and they spend four minutes hunting for the match. They get it right. And then they run out of time and never even read the last three questions, all of which were easy.

Nora Ashford

So they traded three easy points for one hard one and felt productive the whole time.

Adrian Calloway

They felt like they were fighting. Today is about not fighting. Today is the whole machine, how you walk into a thirty-five minute section, and how you turn a stack of missed questions into a higher score.

Adrian Calloway

Quick retrieval first, reaching all the way back to episode one. What are the three parts of every argument on this test? Take a beat before you answer.

Nora Ashford

Conclusion, the main claim. Premises, the support. And the assumption, the unstated bridge across the gap between them.

Adrian Calloway

Here is the thing that makes this the capstone. Every tool in this course is one move performed on that one skeleton. The bridge for assumption and strengthen and weaken. The flaw taxonomy. The three causal attacks. Stay-in-the-box for inference. Abstract-to-the-shape for parallel. The Disagreement Test for point at issue. There are roughly a dozen named question types on this test, but they are not a dozen unrelated skills. They are a dozen ways of poking the same conclusion, premise, gap.

Nora Ashford

So the whole course collapses into one structure with different handles bolted onto it.

Adrian Calloway

And that collapse is exactly what lets you move fast. Here is the objective for today, and notice it is a speed claim, not a knowledge claim. By the end of this episode you will be able to hear a question stem and, in about five seconds, name the tool you are about to run, before you even fight the stimulus. One skill: matching the question to the method on reflex.

Adrian Calloway

Three moves today. Move one, pacing and the skip rule, how you survive the clock. Move two, the decision tree, how you pick a tool from the type. Move three, blind review, how you turn misses into points between practice sets. Let us do pacing first.

Nora Ashford

Give me the numbers. How long do I actually have per question?

Adrian Calloway

A scored Logical Reasoning section is thirty-five minutes with roughly twenty-four to twenty-six questions, usually around twenty-five. There is no fixed count, so do not anchor on an exact number. Divide it out and you land near one minute twenty per question. But here is where most people go wrong. They take that number and ration their time equally.

Nora Ashford

I mean, that is the obvious read, right? One twenty per question, so I give every question about one twenty. That is literally what the average says to do. If a question is running long, I am just behind, and I push to catch up.

Adrian Calloway

That is the trap, and I want to catch you in it on purpose, because almost everyone falls in. The average is a budget for the whole section, not a rule for each question. Picture two questions side by side. One is a clean Must Be True you can lock in forty-five seconds. The other is a tricky Assumption that genuinely needs two minutes. Equal would be one twenty each. Smart is forty-five seconds on the first and two full minutes on the second. Same total time, but you bank the surplus from the easy one and pour it into the hard one.

Nora Ashford

Oh. So when a hard question runs long, I am not falling behind, I am spending savings I already earned on the easy ones. The per-question number was never the point.

Adrian Calloway

That is the whole philosophy in one line. Concretely, a widely taught phased plan: clear the first ten or so questions in roughly eleven to twelve minutes, with high accuracy, because the front tends to be friendlier. That banks you around twenty minutes for the harder stretch. And the hard cluster usually sits in the middle to back, often somewhere around questions fifteen to twenty-two, not in a clean climb to the very last one. So protect that back-half time. These are calibration guidelines, by the way, not laws from the test-maker, so tune the exact numbers to your own speed.

Adrian Calloway

Now the move that prevents the four-minute mistake from the cold open. The skip rule. Picture Nora on a mock section. She rips through questions one to ten in eleven minutes, feeling great. Then question sixteen, a Parallel monster. Ninety seconds in, she still cannot get the structure to resolve. What should she do?

Nora Ashford

Honestly, my gut says keep going, I am so close. But what if I could have gotten it? Walking away from a question I almost had feels like quitting.

Adrian Calloway

I want to honor that fear and then dismantle it, because the fear is doing real damage. Here is the rule. At about ninety seconds with no real traction, meaning you still cannot find the conclusion or you cannot eliminate more than one answer, you make your best guess, you flag it, and you move. And the reason the fear is misplaced is opportunity cost. The two more minutes you would spend grinding question sixteen could answer two or three easy questions later in the section. You are not protecting a point. You are spending three to save one.

Nora Ashford

And there is no penalty for a wrong guess, so the flagged question still gets a bubbled answer no matter what.

Adrian Calloway

Never leave one blank. Guess before time expires on every single question. Now, that skip move gives us the two-pass structure. Pass one, you go front to back, you answer everything that comes cleanly, and you flag anything that stalls you using the digital interface flag tool. Pass two, you return to the flagged ones knowing exactly how many minutes are left. And the second look is fast, because the answers you already eliminated stay eliminated. You are resuming, not restarting.

Nora Ashford

So there are two ways to do this badly, and they are opposite. Flag everything and I never finish pass one. Flag nothing and I burn four minutes on one question.

Adrian Calloway

Both ends are failure modes, and that is the subtle part: skipping is not a vibe, it is a threshold. The skip is targeted. Only when you are genuinely stalled at ninety seconds, and always with a guess left behind.

Adrian Calloway

Move two. The decision tree, the spine that ties the whole course together. The idea is simple but it is the entire game: identify what the question wants before you commit to a method, then run the matching tool. Because the type secretly dictates what a right answer is even allowed to be. An inference answer has to stay inside the stimulus. A strengthen answer has to go beyond it. Different jobs, so different tools, so picking the tool blind is picking blind.

Nora Ashford

And you get that type from the question stem, right? Like, the stem tells me it is a Weaken before I have even read the argument, so obviously I should read the stem first.

Adrian Calloway

Let me slow you down, because this is genuinely contested and I will not pretend otherwise. Reading the stem first is a real, defensible approach, and some well-known instructors teach exactly that. But it is not settled. Other major programs teach reading the stimulus first, and they argue that peeking at the stem adds a task to your plate and distracts you from deconstructing the argument. No approach has been shown to produce higher scores. So the honest rule is this: identify the question type as part of your read, and experiment with whether the stem or the stimulus goes first to find your own optimum. The non-negotiable part is that you know the type before you commit to a tool.

Nora Ashford

So I overstated it. The order is a personal-optimization thing I have to test. The identification is the part that is actually mandatory.

Adrian Calloway

That is the correction. So, lightning round, and this is the skill. I read you stems only, no stimulus, and you call the type and the tool. Number one. Which one of the following, if true, most weakens the argument?

Nora Ashford

Weaken. So it keys off the gap. I build the bridge, then I attack it.

Adrian Calloway

Number two. The reasoning is most vulnerable to criticism in that it...

Nora Ashford

Flaw. I run the flaw taxonomy and name the error: necessary versus sufficient, ad hominem, correlation treated as causation, and so on.

Adrian Calloway

Careful, because here is a distinction that trips people for months. A gap and a flaw are not the same animal. In a Necessary Assumption, the gap is something the author needs to be true. In a Flaw, the gap is an actual error in the reasoning. Same crack in the structure, but in one you quietly fill it and in the other you loudly call it out. Blur them and you will start hunting for errors on Assumption questions where there is no error to find.

Nora Ashford

Got it. Assumption, the gap is load-bearing, I supply it. Flaw, the gap is a mistake, I expose it.

Adrian Calloway

Number three. The two speakers are committed to disagreeing over whether...

Nora Ashford

Point at issue. Disagreement Test. I need one proposition where speaker one says yes and speaker two says no, and both positions provable from what they actually said, not just both of them talking about the topic.

Adrian Calloway

And the rest of the map, fast. If it says must be true or most supported, that is inference, stay-in-the-box, the answer has to be one hundred percent provable from the stimulus alone. If it is parallel reasoning, abstract to the shape, match the logical form and never the topic. If the conclusion claims one thing caused another, reach for the causal attacks. The type hands you the tool. That is the entire decision tree.

Adrian Calloway

Let me prove how much the type drives everything with one argument viewed three ways. Picture this. The office coffee machine broke the same week productivity dropped, so someone concludes the broken machine caused the slump. Hold that in your head. Before I tell you the question type, you tell me. Lens one, the stem says most weakens, and the conclusion is causal. Which causal attack fits?

Nora Ashford

Alternate cause. Something else explains the drop. Picture it: that same week there was a brutal deadline crunch that tanked everyone's output, and the dead coffee machine is just a coincidence sitting next to it.

Adrian Calloway

Clean alternate-cause weaken. And a quick honesty note. Alternate cause, reversed causation, and mere coincidence are the three core attacks, but they are a simplification, not the whole list. The fuller taxonomy also includes showing the cause with no effect, the effect with no cause, and attacking the underlying data. Three is your core kit, not the universe. Now lens two. Same stimulus, but the stem says the reasoning is flawed. What is the error?

Nora Ashford

Correlation treated as causation. Two things happened the same week, and the author leapt straight to one causing the other.

Adrian Calloway

Notice the move changed even though the words on the page did not. As a Weaken, you add a brand-new fact, the deadline. As a Flaw, you add nothing at all, you just name the logical sin already sitting in the argument. Now lens three, the contrast that really teaches it. Same coffee-machine argument, but now the stem says which one of the following must be true based on the statements above. Pause and try this one before I let Nora answer. What is a right answer even allowed to be here?

Nora Ashford

Only what is one hundred percent provable from the text. So the safe answer is just, the machine broke and productivity dropped in the same week. That is it. I cannot say the machine caused anything, because the stimulus only handed me the timing.

Adrian Calloway

Say why that step works, out loud.

Nora Ashford

Because inference is stay-in-the-box. The author's causal leap is their opinion, not a proven fact, so a Must Be True answer that repeats the leap is unsupported. And I should be suspicious of any answer with all or never or always unless the text earns it. The softer words, some, may, can, are usually the safer ones.

Adrian Calloway

And do not import the real world. Even if you personally know a broken coffee machine tanks morale, that knowledge is contraband on a Must Be True. One stimulus, three stems, three completely different right answers. The type is the steering wheel, the stimulus is just the road.

Adrian Calloway

Move three, and this is the one that actually raises your score between now and test day. Blind review. Here is the mistake almost everyone makes. They finish a timed section, they immediately flip to the answer key, they see the right answer, they go oh, of course, and they learn almost nothing. Because seeing the answer first is reverse-engineering, not diagnosing. Your brain rationalizes the credited answer instead of revealing why you missed it.

Nora Ashford

So what is the right sequence?

Adrian Calloway

Take the section under strict time first, flagging every question you were unsure of as you go. Then take a real break. Then, before you look at any answer key, you go back to the flagged questions and re-solve them untimed, to one hundred percent certainty. No clock, no key, just you and the argument until you would bet money on your answer. Only then do you check the key.

Nora Ashford

And the payoff is the comparison between my timed answer and my untimed answer.

Adrian Calloway

That comparison is the entire diagnosis, and it splits your misses into two completely different problems. Pause and predict before I give it. Case one, you missed a question timed, picked the tempting trap, but in blind review, with no clock, you calmly got it right. What does that tell you?

Nora Ashford

That I actually knew it. I just ran out of time or panicked under the clock. So that is a timing problem, not a knowledge problem.

Adrian Calloway

A timing miss, fixed with pacing and the skip rule, not with re-studying the concept. Now case two. You miss the question timed, and you still miss it untimed in blind review, with all the time in the world. That is a fundamentals miss. You do not understand the concept yet, and no amount of speed will save you. That one sends you back to the relevant episode and the method, never to a stopwatch.

Nora Ashford

So the exact same wrong bubble can mean two opposite things, and blind review is the only thing that tells them apart. Without it I would just be guessing at which fix to apply, and probably grinding speed on a question I never understood.

Adrian Calloway

That is why it is the engine and not a chore. And you pair it with a wrong answer journal. For every miss, you log it, you reconstruct why the trap answer was tempting and why the credited answer is right, and you watch for patterns by question type. If three weeks of journaling shows you keep dying on Parallel and Point at Issue, your practice stops being random and becomes a targeted hit list. That is how passive re-reading turns into real, repeatable learning.

Adrian Calloway

Here is your one heuristic for the whole section, the thing to carry into the room. Identify, then attack. Read to name the type, then run the matching tool, and never let one question hold the others hostage.

Nora Ashford

Identify, then attack. And the person in the cold open did it exactly backwards. He attacked question sixteen for four minutes and never identified that the last three questions were free.

Adrian Calloway

He fought the wrong battle. Identify, then attack, and skip what stalls you. Now the honest part. This is the final episode, and I need to say plainly what this course is and is not. Everything across these twelve episodes is the review-and-recall layer. We compressed the patterns so they sit at the front of your mind: the skeleton, the bridge, the flaw taxonomy, the causal attacks, stay-in-the-box, abstract-to-the-shape, the Disagreement Test.

Nora Ashford

But listening to us does not build the speed.

Adrian Calloway

It cannot, and no audio can. The pattern-recognition speed this test rewards only comes from reps under time. So here is the concrete next step, and it is the only homework that matters. Today, open a question bank and drill one full, timed Logical Reasoning section. Then run blind review on it: flag, re-solve untimed before the key, sort your misses into timing versus fundamentals, and log them. The audio keeps the patterns warm. The drilling makes them fast.

Nora Ashford

So replay us on the walk to the desk, then do the reps that actually move the score.

Adrian Calloway

That is the deal. You came in seeing a wall of question types. You leave seeing one skeleton and a handful of tools you can name on reflex. Identify, then attack. Now go put it under the clock. You have got the method. Go earn the speed.