Helpful LSAT Prep Tips

Whew. Just the top three? Really? OK, here goes:

1. Think about the way that you've learned to get good at anything else you've done - hard work and smart practice. Expect to put in real effort if you want to succeed at this. Anticipate that you can probably benefit from a good coach. Know that you'll have to do organized review in order to grow.
2. Don't just take practice tests - try instead to gain mastery of each individual question type. This will give shape and focus to your prep. In order to be fast for test day, you'll need to be good at what you do.
3. (Really more of number 2) Pattern recognition is key. For example, there's just not enough time in a Logical Reasoning section (at least not for anybody over here - Dave included!) to analyze every argument. Instead, we've figured out how LR passaged are related to others by kind. That way, you get to recognize answers rather than having to synthesize them anew for every question you work.

OK, we know we cheated; that first one was, like, 6 tips squished together or something. Sorry.

This is an important question, and one that I get a lot. Since it seems to have deep and complex psychological roots, test-day anxiety is tough for me to crack. But here are some thoughts that might help:

1. Knowledge kills fear. If you know that you're ready for everything the test will ask, then you can relax a little, secure in the knowledge that there's nothing they can throw at you that you haven't seen before. Dedicated, focused practice is required to get to that place, but I believe you can do it.

2. Do some of your prep in the food court at the mall. If you can learn to think under those conditions, then you'll be able to deal more easily with the smaller noises and distractions of the other test takers around you on test day.

3. Many prep companies offer free practice exams - take advantage (if you haven't already), just for the experience of sitting in a strange room, surrounded by strangers.

4. Learn and practice some visualization techniques: Picture yourself in the room on test day. In your head, walk through the entire experience of the day, from waiting in line, showing off your passport photo, to receiving the test booklet and listening to the inexorable drone of instructions, on through opening the seal with your pencil and turning the first page. Picture yourself owning that test. If you can see all of that, then once it happens, you'll already be in charge of it.

5. Find yourself a theme song. For my most recent test, I used "No Church in the Wild". You need something that kicks ass—you won't go wrong with "Eye of the Tiger". Play your song before every practice test you take, and play it on repeat as you drive to your test center on the day. YOU ARE ROCKY MOTHERF**ING BALBOA. 

6. Remember that anxiety is an evolutionarily pre-programmed response. It's nature. You cannot expect to not feel nervous, and just saying to yourself "Hey, self, don't be nervous," won't stop it. Instead, recognize that the nerves are part of the experience, and that they will come, but that they do not get to win. The nerves don't mean anything. You are in control, and you will stay in control by allowing the anxiety to come on, recognizing it for the purely physical reaction that it is, and moving straight on past it.

The same way you get faster at anything else!

This may help contextualize the job for you:

The advice below is from one of the best books I've ever read (really. You should absolutely check it out).

It's called Born To Run, and it's about the Tarahumara, an ancient tribe of Mexican Indian runners. All they do is drink (home-brewed) corn beer, wrestle naked, screw (an obvious outcome of the naked wrestling), and run. They can run for days. They're faster, as a group, than anyone in the world. It's unreal - hundreds of miles at a stretch.

The author of the book spent time with them, and when I came upon this lesson, offered by a man who'd spent years with the tribe, I was blown away by how exactly it mirrored my advice for gaining speed on the LSAT (the book is by Christopher McDougall, and I urge anyone reading this to pick it up. It's breathtaking, and suspenseful, and moving, and inspiring). Anyway, here's the passage:

“ ‘Lesson two,' Caballo called. 'Think Easy, Light, Smooth, and Fast. You start with easy, because if that's all you get, that's not so bad.

Then work on light. Make it effortless, like you don't give a shit how high the hill is or how far you've got to go.

When you've practiced that so long that you forget you're practicing, you work on making it smoooooth.

You won't have to worry about the last one—you get those three, and you'll be fast'."

Emphasis mine.

For your Practice Exams, I see at least four equally legitimate options, and I think you should probably use a mix of all four. In the earlier stages of your prep, the learning/review-type approaches are likely better for you. Then, as you gain confidence, you may want to do more practice/dress-rehearsing.

1. Take it for learning. Don't time yourself. Do it with your manuals at the ready, and take it nice and slow, making sure you suck the marrow out of each question. You can even look up right answers question-by-question as you go. If you take this route, it's very important that you never be satisfied that the right answer choice is (B), but that instead you make it your mission to understand why answer choice (B) is correct.

2. Take it like review. One thing I like to do to check that the skills are sharp is stop-time drills. Set the clock to run at 35 min/section, but pause it every time you run into something difficult. Then, write down what's difficult about the problem. Write what you think the answer is and why you think the attractive bad answer(s) is/are wrong. This will give you a real-time assessment of where your head's at, giving you a clearer review process than you could get after the fact.

3. Take it like practice. Time yourself, but with a stop-watch, not a countdown timer. Keep the stop watch face-down, and don't worry about the time (that part's just for helping you know how fast you're actually moving) - instead, concentrate on doing the work to the best of your abilities. At the end of every section, note the time it actually took you to complete the section along with the questions you got right and wrong. This way, you'll get a reading of how much you can do, and a sense of what you need to get better at in order to do it all within your 35-minute limit. (You can see how you could combine options 2 and 3 if you wanted to. You know, for fun).

4. Take it like a dress rehearsal. This method is probably most appropriate to the later part of your studies. If you do it this way (at a neutral site, strict timing, correct break after Section 3), you'll get the most-accurate possible measure of your current scoring range. This can be a useful barometer. If you do this route, then a proper review of your test becomes extra-super important.

You may want to take 5-section (rather than the LSAC-provided 4-section) practice tests to better mimic your test day experience. If so, you can accomplish this quite easily! Here's what we recommend:

1. Set up your 5-section practice tests in four-test cycles. So, for example, set aside PrepTests 51-54 to use as 5-section timed tests.

2. Then, add one more test to that group, which will be your dedicated "experimental" section material. For example, you could add PrepTest 50 for this purpose. Just write EXPERIMENTAL across the top in black marker.

3. Now, each time you take a timed practice test, add one section from your EXPERIMENTAL test at the end.

Presto! You've got a 5-section timed test!

One of the most important things to do in your learning is achieve a studied, deep review of your practice exams - something that can help you catch the points you'll need to hammer down.

Here's the basic approach I recommend:

1. Make sure that you completely understand why every correct answer is correct. I'm always around if you need help, but you have to totally get the reasons for right answers.

2. Make sure you can articulate the weaknesses of any bad answer choices you fell for. Again, I can help if you're stuck, but you need to really understand what makes an attractive bad answer wrong.

3. Think of it this way: If you had to teach someone else how to answer the question, what would you say?

While you're taking a practice test, find a way to mark every question that you're not 100% on (write the question number on your scratch paper, for example).

It's totally cool if you end up with literally every question noted on your scratch paper. Just take time to mark every instance where you're not sure.

Then, when you do your review, concentrate on confirming all the right answers to those marked questions—you made a lot of good choices: Why?—and on getting a deep understanding of where you went wrong. You'll need to know why every right answer is right if you want to repeat good performances and eliminate bad ones.

Couple review of those questions with particular attention paid to any questions that you missed and yet didn't mark (i.e., the ones you thought you were 100% sure of), and you'll have a comprehensive overlook of your areas of weakness, so you can systematically eradicate those weaknesses.

Yes! Slow down.

I say slow down, when what I really mean is relax. The difference, measured in seconds, between the amount of time it takes you to read a sentence:

"As to the issue of Justin Bieber's romantic inclinations, I could care less"

or to correctly read the sentence:

"As to the issue of Justin Bieber's romantic inclinations, I couldn't care less"

almost can't be measured.

Probably you can't measure it using your wristwatch, anyway. But that miniscule difference in time means the difference between an author caring about the Biebs' love life in the first instance, and not caring about the Biebs in the second.

To catch that distance, you just have to read, which shouldn't be hard, but sometimes is hard nonetheless. So my best advice to you is to try to slow down, just a bit. You have the time to read properly. Make yourself know that - where it counts - and your actions will start to reflect it.

I think that will help. Hope it does.

I'd say the best way to think of conditional symbolization is as a tool - really, in pretty much precisely the way you'd think about a hammer.

When you're driving a nail, there's just nothing better for the job than a hammer. I mean, you could do it another way (whacking at it with your tape measure, or with the sharp end of your screwdriver, if you're really precise and awesome with a screwdriver), but for putting a nail in, the hammer is your best bet.

At the same time, however, you don't go running downstairs to your toolbox to fetch your hammer every time you need to push in a thumbtack.

Conditional symbols are like that hammer - you should use them when you find they help you see the relationships, and you should feel fine with not using them if you can decipher the relationship in the passage easily without them.

As to language cues that tell you that you've got a conditional relationship, I'll direct you to this document, which contains all the conditional language in the world* (*probably not, really, but I am updating it every time I come across conditional language that isn't already on it. So, you know…).

My, what sharp eyes you have!

You are absolutely correct in your assessment that sometimes, I act as though I seem to think "most" means the same thing as "all" (or perhaps you've found me treating some other non-conditional relationship as though it were conditional)

And you're also correct in saying that those phrases are different. And you're also right to say that that difference is important.

So let me explain myself:

My use of conditional symbols for statements that aren't actually conditional is a kind of approximation.

Like this:

Imagine a multiple-choice math question that asked you for the product of pi and 13.

I know that pi is basically 3, so I can estimate the answer before I have to use my calculator to do any actual math. So, here, I would begin my assessment of answer choices by seeking an option that says 40 (give or take).

Now, pi is not 3. It's similar to 3, but it is absolutely different in a real and important way. Nevertheless, I may find great utility in beginning my approach with some approximation.

To continue the metaphor, imagine that the answers are:

(A) 22
(B) 27.5
(C) 31.45
(D) 40.84
(E) 49

I've got the right answer, even though I did so using a value that's inherently, really different from pi - and I did it fast.

Of course, what if the answers look like this?

(A) 39.85
(B) 40.15
(C) 40.45
(D) 40.84
(E) 41.84

In this instance, obviously, my approximation does me no good at all, and I'm going to have to be more precise in my calculations.

Questions in the LR are just like this.

I think it's appropriate and useful to begin our approach to a particular question by making use of estimation and approximation when we can (because estimates are often faster, simpler, and easier to perform than actual calculations), but that we must be prepared for the possibility that our estimation may prove insufficient for answers.

When that happens, we tighten up our calculations.

Yeah, "the only" is the tricky one.

As we've discussed, only and only if are arrows.

The only, on the other hand, means the same thing as All.

So, Only an idiot would buy that record, which clearly means the same thing as Buy that record only if you're an idiot, is symbolized the same way:

Record → Idiot

However, those are different from The only idiots here are the people who bought that record, which is the same as saying All idiots here bought that record:

Idiot → Record

All of this to say just remember that the only means all and you'll be fine.

The same way you'd get faster at typing (or running, or knitting, or doing Kung Fu forms).

Speed comes on this test from the same place it comes in every other thing we do in the world - by getting better at it.

Recognizing patterns is important (probably even essential - I know I couldn't've gotten all the questions right in the time allotted if I'd had to analyze each question individually). Instead, I learned to identify things by type. I sought patterns within the sections.

When I talk about pattern recognition, what I mean is this: something like 80-90% of the arguments within any Logical Reasoning section share something in common with other arguments in the section. To become a faster, fiercer test taker, you should spend a tremendous amount of time identifying those patterns, so that when you see an example of an argument type you've seen before, you can answer the question through recognition rather than analysis.

This work of identification and classification was the single most important thing I did in learning to succeed on this test.

Fortunately, time and accuracy are essentially two ways of measuring the same thing: how good you are at LSAT.

So, to improve them both, I recommend that you work in tandem on the following skills:

1. Pattern recognition. (that link will synthesize some of the ideas you've already seen in your course videos)

2. Language cues.

3. Slowing down! It helps you avoid careless errors, which helps you earn more points (both by not missing easier questions and by not spending too much time dealing with easier questions; if you read clearly and accurately the first time, your work will be more efficient and thus also faster).

4. And always do your homework according to your study plan. It'll help you organize and manage all the above.

Start here:

Part (A)

Go through an entire LR section, looking at each question. For each one, answer the following questions:

1. What exactly does the question demand from you? (You must be able to answer for 26 of the 26 questions)

2. What can you expect the right answer to do? (You must be able to answer for 26 of the 26 questions)

3. What can you expect the right answer to sound like? (You must be able to answer for at least 22 of the 26 questions)

4. What is one wrong answer likely to say? (You must be able to answer for 26 of the 26 questions)

Part (B)

Then, go back through that section (without bothering with questions this time) and read each passage. For each, answer the following questions:

1. What is the main conclusion of the argument? (You must be able to answer for at least 20 of the 26 passages - no main conclusion for Inference and Resolution Questions!)

2. What is wrong with the argument? (You must be able to answer for at least 16 of the 26 passages - several passages don't exhibit flawed reasoning)

3. How is this flaw typical of others you've seen? (You must be able to answer for at least 10 of the 16-19 passages that exhibit flaws)

4. Which words are likely to be most important in determining the correct answer (no matter what question you might be asked)? (You must be able to answer for at least 20 of the 26 questions)

If you cannot answer all eight of these questions in the proportions indicated, then you do not yet have a strong enough grasp of the fundamental principles involved, and you'll know what you need to work on!

Totally! Here's three things for you to try:

1. Try to think of the sentences within the argument as pieces of a puzzle. Their only meaning is their utility—they want to fit together so that there's no daylight between them, but they don't. So ask yourself why they don't. What's the missing piece? This is structural reading.

2. Try framing the argument in the following way: "This author believes Thing X because of Things Y and W. So, what do Y or W have to do with X?" This, too, is a way to read structurally.

3. When faced with a typical flaw, name each piece structurally, as a part of that flaw. For example, in 29.1.3, we have both a Prescriptive Error and a False Choice Flaw. I'd say the following: "This author wants us to use more gasohol (that's her prescription), but she never demonstrates that gasohol will be better than gasoline (that's the flaw!). Instead, she offers two benefits (more octane, less CO) and one feature that may not even be better (it doesn't add more CO₂ than plants remove. Does gasoline do that, though?). These three factors present a false choice—they neglect the possibility that there are other, possibly even more important factors!" Then I'd take a quick nap because that's a mouthful. But doing that repeatedly makes me better at it until I go from working it out, to thinking it, to just seeing it. That seeing is structural reading.

OK. Strap yourself in. This one's a long, multi-part answer:

Answer Part 1 - Inference Questions

For inference questions, remember that the right answer is something that you can prove based on the passage, right?

So, think about it this way: If you have four or five lines' worth of passage, how likely is it that in those few lines you'll be able to prove that "Most successful entrepreneurs have engaged in and enjoyed carnal relations with root vegetables"? Not likely at all, right? I mean, to prove it, you'd have to know how many successful entrepreneurs there are in the world, plus how many of them have engaged in the disgusting relationships indicated here. And if a passage had told you that much information, it would be too easy to answer.

Instead, you're much more likely to prove that "At least some successful entrepreneurs have engaged in and enjoyed carnal relations with root vegetables." To so prove, you don't need to know how many successful entrepreneurs there are in the world, and you don't need to know how many of that aggregate have filthy, filthy habits. You'd only need one example in order to prove that "some" enjoy those relations.

What I'm saying is this:

When you've got it down to two choices in an Inference question, choose the one with smaller, softer language every time. Generally, you'll eschew words like most, usually, all, never, and only, in favor of words like some, sometimes, not all, and not always.

Answer Part 2 - Sufficient Assumption Questions

These questions demand of you that you provide an answer that, if true, would be sufficient to prove that the argument's conclusion is true. And proof? That means that it's impossible for that conclusion to be false.

That's a really big job. How do you do that?

I mean, how do you prove, in a sentence, that some claim is true? It would take some heavy-duty information to do that, right?

So, expect that the right answer to a Sufficient Assumption question will be big. Expect it to employ what I call Load-Bearing language (the kind of language that can bear the burden of proof).

Words like all and always and never and every and only. Also superlatives - words like best and first and smartest and weakest and surest.

When choosing between two answer choices for a Sufficient Assumption question, choose the more-aggressively worded choice.

Answer Part 3 - In which we switch gears, to talk Reading Comprehension

In the RC, every question asks you what is supported "according to the passage," right? That means, of course, that the right answer to every question in the RC can be found within the passage that's on the page next to you. It tells us that the answer is available in print. That's something - it means that if we look efficiently enough, we can be guaranteed of finding the correct answer. I love RC because it's like an Easter-egg hunt in this way. The answers are all right there, just waiting to be uncovered. It's so great! So, that's one thing.

What it also says for us, that we may overlook, is that all RC questions are Inference questions. Every one asks us - explicitly or in essence - what we can prove on the basis of the passage. This means that the answers to all RC questions are Inference answers - the right answer choice, then, will tend very strongly to be small.

So, for all RC questions, when choosing between two answer choices, choose the smaller of the two. Exactly the way we talked about Inference questions in Part 1.

One other thing:

Imagine two answer choices, when you've been asked for the author's attitude. (A) says "scornful" and (D) says "critical." I can tell you right now, without any passage to reference, that the correct answer between those two has to be (D). I don't need any evidence at all in order to be positive in my choice.

Here's why: if an author is "scornful," then she must also be "critical." It isn't possible to express scorn without the element of criticism. So it isn't possible for "scornful" to be correct in this instance - if (A) were true, then (D) would also have to be true. And it's not possible to have two correct answers. On the other hand, it's entirely possible to be critical of someone without being scornful of her. There's no reason that (D) can't be true without needing (A).

In RC, when in doubt, choose the smaller answer.

You'll be right most of the time.

Answer Part 4 - Method Questions

Method questions ask you to articulate the method of reasoning used by the argument. The demand can be suitably met if you think of every Method question as asking you this: "Please describe the argument above" (in which the introduction of the word "please" makes the whole thing go down better, don't you think?).

So you begin your answer by describing the way the argument goes in gross, structural terms. And then you find an answer that matches that description. But what if there are a couple that seem enticing?

Before we tackle that directly, here's something important to keep in mind when you're looking at answer choices: the test writers are very unlikely to take sides. That is, the writers do not tend to denote any agreement with (nor disapproval of) the content of these passages. They just put the information out for us to reason through. And this makes sense, right? I mean, if the test writers were to indicate that any conclusion was, in fact correctly drawn, then that could potentially lead to challenges (this whole test is, after all, being administered to a bunch of future lawyers).

So rather than take any part, the test writers will pepper their work with phrases like "...according to the passage..." and "...if the above statements are true..." at least partly so that they, as an entity, cannot be accused of having taken an invalid or unsound - or just unpopular - stand on any issue.

Now, here's how we can use that bit of knowledge to our advantage: The right answer to a Method question is very unlikely to be declarative. Instead, the right answer is much, much, more likely to be qualified in some way. So, for a question asking

"The argument proceeds by..."

between these two answer choices:

(A) undermining a claim by attacking the evidence offered in its support

and

(D) attempting to refute a view by showing that it rests upon a questionable assumption

choose (D) every time, even if you don't have a passage in front of you to compare these answers to!

Here's why: if (A) were credited, that would mean that the test writers have said that this argument has in fact undermined some claim. This means the argument worked. Now, it's possible to undermine a claim, but it's incredibly unlikely that the test writers would ever indicate that an author has done so successfully (as we discussed above).

Instead, it's much more likely that the test writers will indicate that the argument has proceeded by attempting to refute a view. See, this removes all voice - the test writers aren't saying that the argument succeeded, and they're not saying that it didn't succeed. They're just describing - saying that the argument tried to refute a view.

And that's the kind of carefully-worded answer choice that's likely to be correct.

So, for Method answer choices, choose a moderated, qualified statement over an unmodified declarative answer choice. Almost all the time, you'll be correct.

Answer Part 5 - Necessary Assumption Questions

The right answer to the Necessary Assumption is, well... necessary to the conclusion. It's a piece of evidence that the argument needed, but left out.

This means two things for us:

1. The right answer will often be small. Like most other necessary things in the world, the right answer is not likely to be something big and aggressive. Consider the things we need in life. We don't need the $600 bottle of Cristal, we don't need the McMansion - or the McDonald's, for that matter - we need only some food, some shelter, something to drink. In the same way that our physical needs tend most often to be little, the right answer to the Necessary Assumption question will tend to be little. Expect often to see the word "some" or "not all" and the like.
2. The right answer is necessary - meaning essential, required, un-live-without-able, right? In other words, if you take it away, the argument will die without it. So, when you think you've found the right answer, ask yourself this question: "If this answer choice weren't true, would the conclusion still make sense?" If the conclusion can live without the answer choice, it's not the right answer! If the right answer isn't true, then the conclusion of the argument will become stupid. This is because the conclusion depended on the truth of the right answer.

So, when choosing between two answer choices in Necessary Assumption questions, choose the smaller answer, and try the test discussed in point 2 above.

Answer Part 6 - Weaken + Strengthen Questions

We've talked a lot about language cues, because they are among the most important indicators of correctness (and incorrectness) when you've got two answer choices left.

There are pretty strong cues at work for Weaken and Strengthen questions, too. Like this:

1. We know that the right answer to a Weaken or Strengthen question will (respectively) attack or assert the necessary assumption of the argument. In order to do so effectively, that answer needs to employ bigger, more powerful words than the soft language we expect from Inference and Necessary Assumption answers.

2. While the load-bearing language we know to expect from the answers to Sufficient Assumption questions would be awesome for doing the work of weakening and strengthening, experience teaches us not to expect such language to be used.

So, while it would certainly kick a lot of ass (in more ways than one) if an answer choice said "Nobody has ever enjoyed any Steven Seagal movie," such an answer isn't to be expected among the answer choices.

At the same time, an answer that said "People do not always enjoy Steven Seagal movies" is such weak sauce that it can't be expected to effectively counter (or support) much of anything, and would therefore be very unlikely to be the credited response.

Instead, we will expect language that's in between those two poles.

I'm thinking about words like these:

• most
• many
• often
• usually
• rarely
• few

I call these words (quite cleverly, I think) "Middle Language."

Denotatively, they're not so different from the soft language of previous discussions. However, these words trade heavily in connotation. So much of their force comes not from direct meaning, but instead from their evocative nature. "Many" doesn't mean much different from "some," but it sure sounds like more, doesn't it? For this reason (along with those given above) we can expect this kind of language within the right answers to weaken and strengthen questions.

So, when choosing between two answer choices for Weaken and Strengthen questions, look for Middle Language.

Answer Part 7 - Parallel Questions

The Parallel Question asks you to describe an argument (exactly the same task demanded by the Method Question), but then, instead of answer choices articulating that description, you're given 5 other arguments, one of which answers to the same description as the original.

There are three quick checks you can perform to get rid of many bad answers to these questions:

1. Does the conclusion match? If our argument was prescriptive, then the right answer must also include a prescription (look for the words should, ought, must, + needs to). If our argument concluded that Thing X was not the cause of Thing Z, then the right answer must likewise conclude that Thing F was not the cause of Thing M. If the conclusion doesn't match in its form, that's not the right answer.
2. Do the quantifiers match? If our argument said "most lizards are herbivores," then the right answer must say "most musicians are bassists," and not "some moles are cancerous." If our argument says "all members of the Kardashian family are vapid pretenders," the right answer cannot say that "many exotic birds are preening narcissists." The language of quantity is a major indicator of the underlying logical structure. It has to match.
3. Do the topics match? If so, that's almost certainly the wrong answer. It is an extremely common red herring used by the test writers to present you with an argument about travel by train, and then a wrong answer choice about travel by bus, in the hopes that you'll mistake a similarity of content for a similarity of logical structure. Don't fall for it. Answer choices with similar subject matter to the argument are almost always wrong (really, I think they've actually always been wrong, but it's certainly possible to construct an argument with the same structure about the same content, so let's content ourselves with a strong warning against same-content answers, rather than a strict prohibition against them).

Running checks of those three items can help shorten the time it takes you to answer Parallel Questions, and can help you decide when you find you have two contenders left among answer choices.

Answer Part 8 - Flaw Questions

There are three kinds of answer choices associated with Flaw Questions. None of the three is more (or less) likely to be correct than the others. Instead, it's just that there are three flavors used by the test writers. Understanding these types of answer choices can help you understand when a choice is wrong and why it is.

Here are the three types:

1. Choices that identify flaws. These answer choices are generic in language - they're not specifically tied to the content of the argument. They describe the flaw in structural terms (see this PDF - the same one I posted earlier - for examples of how LSAC identifies flaws in this way). When dealing with answer choices of this type, you must ask yourself only, Did we? "Did we, in this argument, really [generalize based on a sample, etc.]?" What sample? If you cannot identify the sample, then the argument didn't commit a sampling flaw, and this answer choice isn't correct.
2. Choices that identify the assumption of the argument. These answers most typically begin with these words:

Presumes without providing justification that...
Takes for granted that...

Those two phrases mean "assumes." If an answer starts with that language, then whatever follows must be a necessary assumption of the argument, or it isn't the right answer. Answer choices that begin this way are Necessary Assumption answer choices, and are therefore subject to the same rules and tests as the answers to any Necessary Assumption Question (see Part 5 for more on this topic).

3. Choices that negate the assumption of the argument. These answers most typically begin with these words:

Overlooks the possibility that...
Fails to consider that...
Ignores the possibility that...
Neglects the fact that...

All of those phrases are code for an answer that says "the assumption isn't true." Like this:

The only thing you can overlook (that it is a flaw to overlook) is the dreadful, ruinous possibility that the assumption of your argument is false. Consider this example:

She can't be a supermodel. She's only 5'3".

This argument has assumed that you have to be taller than 5'3" to be a supermodel.

It's flawed because it presumes without providing justification that you have to be taller than 5'3" to be a supermodel.

Another way of saying that is to say that it's flawed because it overlooks the possibility that you don't have to be taller than 5'3" to be a supermodel.

In other words, the reason it's flawed is that it has overlooked the possibility that its assumption is false. The wonderful thing about this 3rd type of answer choice is the clear effect that such a negation has on the argument.

If the right answer to a flaw question begins with this 3rd set of phrases, then whatever comes after them must ruin the conclusion, as in our example:

If you don't have to be taller than 5'3" to be a supermodel, then this argument is blown up. The conclusion becomes stupid, because the answer negates an assumption that the argument was depending on.

So, understand these three types of answers, and you'll go a long distance toward knowing how to get rid of bad answer choices in Flaw Questions.

Answer Part 9 - Point of Disagreement Questions

These questions ask you for the "point at issue" between two speakers, or for the thing about which two speakers "disagree" (hence my choice of name).

When deciding on the right answer between two choices, it will help to have a mechanism for making your judgments.

Here's a three-step approach that may help:

1. First, treat the question exactly like a Main Point question. The thing they disagree about ought to be their central conclusion, right? Articulate the end-of-the-day difference between the two arguers. Look for an answer choice that expresses that point. Experience tells me you'll find such a choice in roughly 1 of every 3 POD questions. When you don't find such an answer choice...

2. If no answer choice describes the main point of contention, treat the answer choices like the answers to Inference Questions; the answer will be the one choice that you can prove the
arguers take opposing positions on.

3. You may find it helpful to draw a two-column T-chart just to the left of the answer choices - one column is for Speaker 1, and the other for Speaker 2. As you read each answer choice, write down in your T-chart Yes or No answers for this question: "Can I prove that Speaker 1 agrees with this answer choice? Can I prove that Speaker 2 does?" If you cannot prove it either way, it's not the right answer. If they both say "no" to an answer choice (or, of course, if they both say "yes"), that's not the right answer. The right answer is the one with a firm Y in one column and a solid N in the other.

The right answer to the POD Question is the think you can prove they disagree about. For that reason, the answer is often pretty aggressively worded. When in doubt, choose the more aggressive answer choice. More often than not, that'll be the right choice.

Answer Part 10 - Main Conclusion Questions

I realize that for this question type, we'll really be talking less about choosing between two answer choices, and more about sussing the right answer from within the passage to begin with. I thank you for your forbearance.

Let's talk about three things that maybe don't get mentioned enough:

1. The But. Probably the single most important indicator of the main conclusion of an argument (in those cases when you've been asked to identify the conclusion, that is). So many arguments take one of these two forms:

(A) But ... since

Some people say that Tony will probably die alone [note the ascription!]

but this can't be true [see the conclusion here?]

since the ladies go crazy for guys who wear lots of cologne and say "broseph" a lot. [since always introduces evidence!]

and

(B) But ... so

Tony is afraid he will die alone, a victim of his own cologne [the setup]

but he needs to remember that there are some ladies who love them some cologne, [ah! competing evidence is offered!]

so maybe Tony will find love someday. [and we finish with a heartwarming conclusion]

In both instances, the word but can serve as an important indicator that you are nearing the conclusion. When you see that word in a phrase, determine whether the following phrase is evidence (as in the but ... since construction) or whether the phrase itself is the evidence (as in the but ... so construction).

2. The After All. These words (just like the word since, and the words to see this and this is because) always precede evidence. As a result of this rule, these words should then function for you like a giant, flashing neon light that says "Excuse me, Carl, but you have just now passed the Main Conclusion of this argument" (I'm assuming that if you're reading this, your name is Carl). When you see any of these words, you know you've found the conclusion. It's whatever sentence you just finished reading.

3. Therefore is wrong. Look, almost all the time, the word therefore is used to identify the main conclusion of an argument. But you tell me - if the test writers have asked you in a question to identify the conclusion, how likely do you think it is that they've telegraphed that conclusion by using the word thus or hence or therefore to signal it? Right. Not at all likely. When you're asked to find the conclusion, avoid the sentence that starts with therefore!

So, for Main Conclusion Questions, you'll be able to find the right answer most of the time (like 80% or better) by following one or more of these three rules:

1. Find the but.

2. Precede the after all.

3. Avoid the therefore.

Answer Part 11 - Role Questions

The Role Question asks you for the function of a statement within the passage as a whole. When deciding between answer choices here, the most important thing you'll do is develop a good initial assessment of the function of the statement.

To do this, start with a wide view - what's the argument about? Describe it in the same way you'd describe an argument when answering a Method question. Perhaps the most important thing to warn you of is to avoid the (completely natural) tendency to focus on the statement in question while you're reading the argument.

If your job is to assess the role of the statement within the passage (and it is!), then you have to be able to see the whole passage in structural terms. If you concentrate on finding that one sentence you've been asked about, you may be less able to provide a good description of the argument, and consequently less able to identify the function of the sentence.

So, for Role Questions, read structurally - the same way you would for Method questions. Prepare to describe the passage as a whole, and only then should you worry about the single sentence you've been asked about.

For answer choices, expect bad answers of these kinds:

1. Choices that correctly identify roles played in the argument by other statements. Your surest way to eliminate these answers quickly is to have at hand a strong, simple assessment of the role of the statement before you look at answer choices!

2. Choices that correctly identify the role of the statement, but incorrectly identify the conclusion of the argument. Many answers will have this form: "Evidence for the conclusion that [X]." Of course, those answers are correct when the statement in question is evidence, and the conclusion of the argument is [X]. However, it's often the case that the statement in question is evidence, but the conclusion of the argument is [Z], not [X]. The existence of these kinds of bad answers (and there are plenty of them) is perhaps the clearest reason why you need to strive for a good general description of the argument, and not worry only about the one statement in question. Stay on your toes, and you'll mostly be fine.

Knowing what you can expect from bad answers will help fill in the puzzle: For Role Questions, you describe the argument, eliminate choices that don't match your description, and eliminate options that don't match the argument.

Answer Part 12 - A final note on language cues

Think of language cues within answer choices sort of like modes of dress:

For example, if a person is wearing a wedding dress, then there are two additional things that are almost certain to be true:

1. The person is female.

2. The person is getting married.

Likely, well over 90% of all cases of persons wearing wedding dresses involve both of the above facts.

Of course, it's certainly possible that one or both of those things are false - maybe you see someone in a wedding dress and he's a man going to a costume party (or maybe he's in New York and he is getting married!)

It's possible for the cues regarding expected language to be misleading, in almost precisely the way that it's possible for someone in a wedding dress to be misleading. It could happen, but not very often.

To review:

For answers to Necessary Assumption and Inference Questions, we want Soft Language.
For answers to Weaken and Strengthen Questions, we expect Middle Language.
For answers to Sufficient Assumption Questions, we want big, burly Load-Bearing Language.

Yes! Let's start with these three:

1. The necessary assumption is a piece of evidence that the argument needs but does not have. The right answer has to be something that fills in a blank within the argument - some place in which the arguer didn't provide you with the evidence, but instead just assumed that the evidence was true.
2. To identify that place, look for shifts in language: Is there some place in the argument where the author changes the subject? If the author provides evidence about one thing, and a conclusion about something else, that author has assumed there's a connection between those two things. That shift in language gives evidence of the shift in logic.
3. The right answer is necessary. That means that if you take it away, the argument will die, right? When you think you've found the right answer, ask yourself this question: "If this answer choice weren't true, would the conclusion still make sense?" If the conclusion can live without the answer choice, it's not the right answer! If the right answer isn't true, then the conclusion of the argument will become stupid. This is because the conclusion depended on the truth of the right answer.

Here's a start:

The flaw of any argument is the fact that the argument has assumed some information. In order to succeed, an argument must move smoothly, building from one point to the next without gap or interruption. When an argument fails to provide sufficient evidence for its conclusion - when it assumes that some important piece of evidence is true rather than demonstrating that it's true - that argument has failed.

Often, an argument will indicate its flaw on the basis of a shift in language: If an argument begins by saying that Mechanical Engineering majors are astonishingly physically attractive as a group, and that therefore, they must be a successful dating population, then that argument is flawed. The flaw is that it has failed to consider that the physical attractiveness of a group may not indicate its dating prowess. One necessary assumption of this argument is the assumption that they physical attractiveness of a group has some relationship to that group's success in securing dates.

So, identify the shift in language, and you'll have found the shift in logic. That shift is where the assumption lives, and that assumed evidence is the flaw of the argument.

Did you really think I'd say no?

Strengthen and Weaken questions are two sides of a coin. In both instances, we will answer by appeal to the assumption of the argument. You cannot strengthen an argument on this test by showing that the evidence on offer is true. We will correctly stipulate the truth of all the evidence (we have to. In a world where the facts are in question, how can we ever hope to reason properly together? See the US Congress for illustration). So, if the validity of the evidence is not in question, how can we make the argument stronger or weaker?

We can do so because the argument has assumed something. So, to make the argument stronger, we'll assert that the necessary assumption is true. To weaken it, we'll deny the truth of the assumption.

Consider an example: If an argument begins by saying that Mechanical Engineering majors are astonishingly physically attractive as a group, and that therefore, they must be a successful dating population, then that argument is flawed, because it has assumed a connection between attractiveness and success in dating.

So, to strengthen that conclusion, indicate that physical attractive does matter in dating. Say something like "Typically, the more attractive a person is, the more likely it is she'll be able to get a date." This doesn't prove that the conclusion is true, but it does make it more likely - and that's what we were asked to do.

To weaken this conclusion, attack the assumption: Say something like "Recent studies have indicated that physical attractiveness is a much less important consideration in dating than financial acumen." In this way, you're denying the strength of the connection between attractiveness and dating. This doesn't prove the conclusion is false, but it makes it less likely. That was its job.

Yes! Here's the basic idea:

I start by describing the argument, trying to take all of the content out of my description.

My description will sound something like this: "That goal cannot be the real reason the person took that action, because there's a better way to reach that goal." (this was my description of PT 37 Sec4 Q 14, if you're keeping score at home).

Then, I skim over answer choices, ignoring those that don't seem to match my description, and spending any analysis time looking at any answer that does match that description.

I'm kind of looking for the forest more than looking at particular trees. This bird's-eye-view strategy takes time to perfect, but it makes you fast when you get good at it.

To test this approach, see if that description doesn't make short work of the example question referenced above.

It's because I think the only useful categorization is one that helps you learn to see things by type. And there isn't a type of "Principle Question." There are four distinct ways that the test writers construct questions with the word "principle" in them, and those four ways correspond readily to four existing types of questions.

Maybe an analogy would be useful here: 

Imagine I wanted to create a taxonomy of world religions. I'd group them by the core beliefs that their adherents share, right? And we'd have a list that went like, Islam, Christanity, Buddhism, Jainism, and so forth.

Now imagine I happen to have four different friends, all named Kevin. Kevin S. is a Christian. Kevin M. follows Islam, Kevin P. practices Jain, and Kevin R. is a Buddhist.

The fact that I know all these Kevins wouldn't lead me to change my religious taxonomy to now say we have Islam, Christanity, Buddhism, Jainism, and also Kevin.

Because "Kevin" is not a type of religion! It's just a name shared by four different people, each one of whom can be usefully grouped within one of our existing categories.

Same thing here.

I'm so glad you asked: as it happens, I do!

The most important thing you can do is find a good way to graphically display the test material. The idea here is that you'll find it easier to see answers than to try to think up answers.

I don't believe I've ever made a deduction while working a game on an actual test. I wouldn't put any credence in a system that requires you to make deductions in games in order to be successful. You can learn to generate and process information visually, and then simply do the pencil work. My method isn't sexy - it's just brutally effective.

Maintaining a visual organization is a low-order thinking skill that you can easily master with practice. Once you have a strong technique, time is not a problem any more. It really doesn't matter precisely how you display information, but it does need to be visual, visual, visual.

(Btw, if you've never taken a practice LSAT, the June test from 2007 is available free here. You also might find it helpful to watch the free video explanations for every question from every recent test available here).

What you can do is seek similarities between games. Virtually every game administered by LSAC this millennium has had strong organizational ties to many other games (and clear cues to those ties). If you can see it, you can work it.

The general principle is this: The games section heavily rewards your ability to efficiently follow directions. The rules of the game are those directions. So, if you can create a strongly visual iteration of the abstractions offered in the setup and rule-set, then you can use your pencil to work out what the rules dictate.

The difference between a deduction-heavy approach and my approach is almost exactly the same as the difference between dividing 7,654 by 12 in your head, and doing long division. The former is elegant and impressive and really hard to do.

I do long division. It's exacting and effective (though not very impressive at all), and it's very easy to do. I have a series of small, simple steps that I perform in the same way, every time, that transform the giant global conundrum of the game into a sequence of low-order actions.

I don't have to think very much during a games section, in much the same way that you don't have to think too hard to do long division with a pencil.

The same way you'd get faster at any repetitive endeavor: Practice!

The best approach to Games is almost exactly like the one you'd use in learning to play the piano. If you wanted to get good at playing the piano, you wouldn't try to do it by getting a whole lot of sheet music and playing each song through once, right? Same thing here.

If you wanted to play piano, you'd learn the finger positions, then you'd play a few songs over and over and over until you could play them fluidly. That sort of mechanical familiarity does two things that are the same things you want to accomplish in the Games section:

• It establishes a skill set. You want to be fast at Games? Learn the process of doing games. You'll learn that process by focused repetition, in which you think long and hard about the most efficient way to deal with the material (and in class, I'll show you some very specific methods for increasing your efficiency). You learn how to do a few things well, and you'll be able to do new and similar things well, too.
• It builds muscle memory - doing the same procedure over and over makes you better - and faster! - at that procedure. Today's games are all so similar to each other - in the same sorts of ways that many piano pieces are similar to each other - that if you learn a few basic moves, those will translate into an almost precognitive action plan on test day. You won't have to think about what you're supposed to do next, if you've already done the thing a hundred times before. You'll just do what comes next, without spending any time wondering or contemplating.

Games is a procedural enterprise - to be faster at the process, you'll want to get more comfortable at performing the series of small manageable steps that go into working any game. Learning the procedure is at the heart of efficiency, and efficiency means speed.

If you are having a hard time picking the "right" template, here's what I think you should do: start every game by just choosing a template. Pick one at random at first, if you need to.

Then, before moving forward, check out the template I used for the game.

Take note of the reasoning I used (in the video) to decide on that template.

Write it down.

Over time, you'll begin to see the repeated features of games that suggest a template.

After even more time, you'll become so comfortable using the templates that you'll feel confident adapting them to the situation in front of you without regard to which template I'd use.

Make sure you're doing the rest of the work according to your study plan, and you'll get better!

Well, in the general, it's not really a question of knowing; there is no "right"—or "wrong"—template to use. Only perhaps a "more efficient" and a "less efficient."

My decisions in every game are informed by repetitive practice aimed specifically at pattern-intelligence, and I thought "Señor Danger [which is what I call myself in moments like these], I think this is the most appropriate template here."

You are now in the beginning stages of that repetitive practice. Keep watching, keep working according to your study plan, and pay attention to the choices you make. Which choices work for you—and which don't—will inform your future choices, based on your experience (and—thanks to your Theory videos—on my experience!).

What’s the difference between a question that says:

1. “Which of the following is a complete and accurate list of all of the cars, any one of which could be parked in the third garage?”

and one that says:

2. “Which of the following could be a complete and accurate list of the cars parked in the third garage?”

Answer:

Question 1 above asks for a global accounting of all of the cars that could ever park in the third garage.

Question 2 asks you for a single possible scenario - of all the possible combinations of cars, Question 2 just wants one such possibility.

You can quickly and accurately assess the different demands posed by these two questions by tracking the position of the word “could” within the sentence:

If the word “could” is near the end of the sentence (as in Question 1), then you’re looking for an answer giving a complete list of all possible entities.

If the word “could” is near the front of the sentence (as in Question 2), then you want a snapshot - an answer choice that gives you one possible permutation of elements to spaces.

To recap:

If could is near the end, then you want a global account.

If could is near the beginning, then you want a single permutation.

I'm totally against the idea of "mastering scenarios". I just hate it to death. A lot, this is because the person who first introduced me to games did this, and it always seemed show-offy and of little help to someone (like me) who wasn't naturally good at doing games. I can totally see how if someone's already really good at games, that itemizing every possibility may seem comforting, but here are the three reasons I never do it, and I don't suggest for my students to, either:

1. If a person can't succeed in a game by following the explicit instructions of the rules, how could one possibly be more successful by isolating and acting upon the implicit inferences it contains? This doesn't make sense to me. It seems like telling someone who's struggling with Algebra to try doing Calculus instead.

2. How do you know when it's helpful to "master scenarios"? There are definitely, looking backward, many games in which there are only four or five total possible outcomes. This is what my trainer talked about, and it always felt ad hoc and hindsight-based. If there's such a thing as some games for which it's useful, and others for which it's not, how do I tell the difference? Again, it seems like in order to tell when it's useful, you'd have to already know how to do the game. If you are instead trying to figure out how to do the game, then adding the question of "Should I 'master scenarios' here?" seems to further complicate an already difficult task.

3. How does it save any time? And here, let’s consider the very typical Game 1 from PT 58. In completing this game, I answered question 1 first, then did 6, 2 and 3. Question 1 is just reading rules, and the other three meant applying rules. My work from those questions answered questions 4 and 5 completely, so in four lines of work (one line of which was just writing down the correct answer from Question 1), the entire game is finished. Looking at this work, I can see that there are at least 10 different ways these monuments could be ordered. So writing out 10 ways would be much, much less time-saving than writing out 4 answers.

That's how I do: It's brutal, but effective for someone who doesn't think in Games, which takes care of point 1. The technique I've described here is precisely the method I use for every game, with the same kind of results, and this solves point 2. Finally, it's much more efficient than any other method - by only answering the questions that I'm asked, I'll never end up with work that I don't need, and my work will always be done in service of directly earning points. This takes care of point 3.

But that’s like, just my opinion, man.

But of course, mon petit fromage.

For Reading Comp, generally, it's the answer choices, not the questions, that make the tasks hard. Answers are written deliberately to seem attractive when they're wrong, and to look ugly when they're right. You can go a long way toward short-circuiting those traps by disciplining yourself to always answer the question based on the passage before you look at any answer choices.

This does two things:

1. Forces you to learn how to properly answer questions (you'll have to learn to stop relying on answer choices, and instead work from the passage itself).
2. Makes you significantly faster over the long run. The place most people waste the most time is weighing answer choices. If you already know what the passage says on the matter, your choice will usually be faster (and more accurate!).

To accomplish these things, get some 3x5 notecards and use them to cover the answer choices. As you work a passage, instead of choosing an answer choice, write down on the card what the passage indicates is the right answer. Only once you've answered each question in the passage this way can you lift the card and choose the answer choice that matches your answer.

I can see a clear straight path from getting good at answering RC questions to being fast at it. It's much harder for me to visualize an avenue for success that doesn't include a disciplined approach to getting questions right.

So glad you asked.

Here are three methods (you should use them all, at least a little bit):

1. Backsolving. Start with the right answer choice for the Main Point and Primary Purpose and Organization question/s (where present). Now you know what the test writers think of the passage: use that to find where the test writers got their support from the passage itself. Think about why they wrote the right answer they way they did. think about how they tried to be tricky in writing wrong answers. This will help you become a more structural reader, but it will also help you develop some insight into the test writers' frames of reference about things!

2. Mimicry. In every RC practice video, I begin by discussing the SPAM of the passage, with some brief notes about anything else I found interesting (I think this is true without fail; you can let me know if you find any place where I didn't do those things). Do what I do! Watch me do it a few times, then do it on your own and compare your work with mine. What did I find important that you left out? What did focus on that I omitted as not big-picture helpful? Whydid I leave that thing out? Now, in addition to thinking like the test writers, you're learning to think like a great test-taker.

3. Finally, think about how you'd teach someone else: what would you say was important for your student to notice about the passage? Take your time with it. Don't worry about your speed—just spend all the time you need to locate the answers within the passage, and think about how each right answer reflects the main point of the passage. Think about whether the passage evinces any of the patterns we've looked at so far. That's structural reading!

I do recommend reading the question first in LR, but not in RC. Here's why:

In LR, many questions do demand the same basic work from you while you read, but there are also important differences between effective reading styles for different question types. For example, in the way you read a Logical Reasoning passage when you're answering an Inference question (in which case, you're thinking about what the passage proves to be true) as opposed to your reading when answering a Necessary Assumption question (in which case you're thinking about what the arguer has left out).

As a result of those differences, I think it makes good sense to read the question first so that you know how to read the passage most effectively.

So, why not in RC?

Theoretically, this system would be just as effective elsewhere as it is in the LR section, but there are two practical considerations that tell me it's not a good system in the Reading Comprehension section:

1. You can't remember 7 questions. The theory is that reading the question first focuses your reading of the passage. The practical truth is that most people simply can't hold the content of 5-8 different questions in their head while at the same time doing an active, engaged reading of a 60-line passage. It's theoretically strong, but practically impossible.
2. You already know what you're going to be asked. In much the same way that you know, for any page of the LR, what 6 or 7 question types are most likely to be on that page, you can know what questions you'll likely be asked on any RC passage. Since you know what's typically asked, you can simply structure your reading accordingly, with an eye to the ideas that are most likely to be tested.

Here's what I do with minimum or maximum number questions (I'll start with minimums, then we'll look at maximums):

I do not try to think about what the minimum might be. I don't try to start from the smallest number. Instead, I start from my prior work, and I do not lose sight of my placeholders.

Here's the process, using PrepTest 34, Game 4, Q 21 as an example: 

1. Eliminate options based on placeholders. Here, for example, we know from our placeholders that we'll never have fewer than 2 doctors at Souderton, so we can get rid of (A) and (B) straightaway. 

2. Use prior work to eliminate other bad options. For example, we've seen as few as 3 doctors in the answer to Q 20, so we can eliminate (E).

3. Finally, start from the smallest number you've ever seen in prior work, then ask whether you can make that number even smaller. For example, if you had tried out answer choice (C) in Q 23 before answering Q 21, then you would know that it's possible to have only two doctors at Souderton. On the other hand, if you hadn't done that question yet (or if you got it without trying out (C)), then you'd start with the smallest number you'd seen before. There are only 3 doctors (N, K, and O) in Q 20, so I'll start from my work for that question.

Because I physically (visually) use my placeholders, I can see that moving K or O out will profit me nothing; I'd have to replace them with P or J, respectively. So if I want fewer than 3 doctors, I need to move N out. If I can do that successfully (i.e., without breaking any rules or adding any more doctors), then I'll know the smallest number possible is 2.

And we can! If we move N out, that takes O out. If O is out, then J is in (so now we have _ J | _ _ N O)

J kicks out K, and when K is out, P is in (giving us P J | K _ N O)

Finally, when P is in, L is out, so we can have as few as two doctors (P and J) at Souderton, while everyone else is at Randsborough.

[Note: if you wanted, you could skip steps 1 and 2 and do the entire question using only step 3. Sometimes, though, you'll get all the way to the right answer by using only steps 1 and 2, so I've included them here because they're not much work to do].

Maximum questions (let's use this same game, and pretend that Q 21 had asked for the maximum number of doctors at Randsborough, and that the test writers had given us options (A) 6; (B) 5; (C) 4; (D) 3; and (E) 2):

1. Eliminate options based on placeholders. Here, for example, we know from our placeholders that we'll always have at least 2 doctors at Souderton, so we can never have more than 4 at Randsborough, and we can get rid of (A) and (B) straightaway.

2. Use prior work to eliminate other bad options. For example, we've seen as many as 3 doctors in the answer to Q 20, so we can eliminate (E).

3. Finally, start from the largest number you've ever seen in prior work, then ask whether you can make that number even bigger. For example, if you had tried out answer choice (C) in Q 23 before answering Q 21, then you would know that it's possible to have only two doctors at Souderton, leaving a maximum of 4 at Randsborough. On the other hand, if you hadn't done that question yet (or if you got it without trying out (C)), then you'd start with the largest number you'd seen before. There are 3 doctors (J, L, and P) in Q 20, so I'll start from my work for that question.

Because I physically (visually) use my placeholders, I can see that moving K or O over to Randsborough will profit me nothing; I'd have to trade P or J to Souderton for them. So if I want more than 3 doctors, I need to move N to Randsborough. If I can do that successfully (i.e., without breaking any rules or losing any more doctors), then I'll know the largest number possible is 4. And, of course, we can move N from Souderton to Randsborough, in exactly the same process we used above.

That's it! That's the process.