How to attack the newest type of LG question? Forum

[nycsoul87]

Does anyone have a decent way to attack the new type of LG question that has been showing up in the last few administrations? It is framed something like this: Which one of the following rules if replacing X rule from the stimulus would allow for the game to be resolved in the same way? Does anyone know what im talking about? I had problems with that type of question in feb so I was just wondering if TLS had a good method of attack.

[09042014]

Good question that shit cost me a 178.

[hellokitty]

No idea but I'd also like to know.

[CryingMonkey]

Although it's not always the case, I've found it can be helpful to brute force this with the aid of the diagrams I've already done. This question will, as I understand it, always come up last. If it doesn't, skip it and do it last. By that point, you should have at least 3-4 solutions diagrammed out from answering other questions. Any of the possible answers that would not be true in the valid solutions you've already found cannot be the credited response. In an ideal situation, only one answer will remain. In the real world, it seems to usually be 2-3. Keep in mind that - so long as the question is framed in this way (which one could replace the rule without changing anything) - you don't need to worry about the replaced rule, since it will essentially still be in effect. With that in mind, I usually look through the responses and try to craft a solution to the game (using the original rules) that breaks as many of the new rules as possible while still being a valid solution. This solution when combined with the solutions I already had usually allows me to eliminate all except for one of the responses. That's your credited response.

As an example, on PT 57, question 5 of the LG section, I knew from previous questions that KJMLHG, KHGJML, JMKHGL, and KHGJML were all possible solutions, with JM and MJ being interchangeable. With that information, I was able to eliminate B, D, and E immediately. From there it's pretty easy to get the answer - JMLKHG is a valid solution, which eliminated A (and also reinforces the elimination of B and E), leaving C as the only possible answer. Double check to make sure that it's true in all of your valid solutions and you're golden.

As I was looking over that one, I realized it also may help to see if any of the answers leap out. Two of the conditions in the original game are M<L and (JM or MJ). Question 5 eliminates M<L, but one of the responses is J<L. This is obviously the credited response, since a simple deductions is that (JM/MJ)<L and switching J<L with M<L doesn't change anything material.

Hope that was at least somewhat coherent.

[d34d9823]

As I was looking over that one, I realized it also may help to see if any of the answers leap out. Two of the conditions in the original game are M<L and (JM or MJ). Question 5 eliminates M<L, but one of the responses is J<L. This is obviously the credited response, since a simple deductions is that (JM/MJ)<L and switching J<L with M<L doesn't change anything material.

This was how I solved it. Just doing the game, you gain quite a bit of intuition that (JM or MJ) implies J=M in any statements relating J/M to other letters. From there, (M<L)=(J<L) is a fairly easy leap.

[CryingMonkey]

Yeah I didn't actually remember doing that question, so I approached it a bit differently than I did when I took the PT. Also it's probalby a bit much to hope that every replacement question will be quite this simple

[BruceBarr]

Good question that shit cost me a 178.

+1 for letting everyone know you got a 178... again...

[09042014]

Good question that shit cost me a 178.

+1 for letting everyone know you got a 178... again...

I got a 176 hence it cost me a 178, but you are welcome.

[d34d9823]

Yeah I didn't actually remember doing that question, so I approached it a bit differently than I did when I took the PT. Also it's probalby a bit much to hope that every replacement question will be quite this simple

Agreed. As a general strategy, I think the easiest route would be to see if you can generate something different given each answer choice. If you're fairly competent, you should be able to eliminate 4 of the answers fairly quickly.

[Richie Tenenbaum]

For PT 57, if you just focused on the deductions in the rules as a result of duplications within the rules, it should not have been a tough question at all. This question needs to be answered last, and make sure you are focusing on what other rules are/can be involved.

If you need to use brute force, that's fine--it's better at just staring at paper. Use previous work and hypos to eliminate answer choices.

These are questions that are harder to prepare for since they are testing understanding of how that specific game works. So there's really no specific prep you can do for this type of question. A good way to prep yourself for this type of thinking though is to go over previously done games and do this exact thing--try to replace one of the rules with a different rule that will accomplish the same thing. This might not be possible on all games, but for a lot of sequencing games this should be possible. Another way to prep for this is to try creating a game; this helps with games in general but it can be a tougher task for a lot of people early on in studying the LSAT.

[NayBoer]

I think I just brute forced it in September. Got -0 on games.

[09042014]

I think I just brute forced it in September. Got -0 on games.

I should have brute forced it. But in my mind, I had a rule that brute forcing meant I was doing it wrong. So I spent time trying to think of a way to do it. I had enough time to brute force it, but I squandered it. I had 15 minutes to the the last LG game, and I didn't finish,

[NayBoer]

I think I just brute forced it in September. Got -0 on games.

I should have brute forced it. But in my mind, I had a rule that brute forcing meant I was doing it wrong. So I spent time trying to think of a way to do it. I had enough time to brute force it, but I squandered it. I had 15 minutes to the the last LG game, and I didn't finish,

I agree, there's always a better way than brute forcing. But I'm a liberal arts major. If I knew how to do math correctly I'd be a geologist for some multinational energy concern, not a 0L contemplating $250k in debt.

[sumus romani]

I recall doing one of these in a practice test and another one on the actual. In both instances, the "replace" question is the last one of the game, and by that point, I had a very good understanding of how the rules worked individually and together. I just solved the question using prior information with a bit of intuition.

[7ED]

Ya. these new "replace" questions are quite brutal.

[theZeigs]

Typically, how many of these questions can be expected on newer PTs?

[LSAT Blog]

On PTs 57, 58, and 59, only one game of the 4 per section had a question like this (I believe the same is true of the undisclosed Feb 2010 exam).

This question appeared as the last question of the game, just as other rule suspension questions always have.

It's likely that you'll see one, and only one, of these questions per LG section in the near future.

-Steve

[kaydish21]

Be careful with these, generally any question that adds or replaces a rule should be the last question you do in the entire LG section, not just on that puzzle. The reason is that this question really could create a whole new question with new rule or logic relations. That said, so far at least, LSAC so far has made all of these questions relatively easy and there has never been more than 1 on a test (not counting my experimental which had 3 and clearly gave away that it was experimental). This is all just rationale, for solving the easiest way is to look for relationships between the new stipulated rule and the rule it replaces. Often there is an overt connection which will only change a minor piece of the puzzle.

[tomwatts]

LSAC so far has made all of these questions relatively easy and there has never been more than 1 on a test (not counting my experimental which had 3 and clearly gave away that it was experimental).

Really? Good lord. I had an experimental games with only one (and yes, there's been precisely one scored version of this question in each test, including February, since it debuted last year).

I hadn't thought of the "use previous work to see if any of the situations you've drawn would be impossible with the answer choices" method. That's an interesting one that I'll have to try. I can't immediately find my notes for 57 and I still haven't gotten around to doing 59 yet, but I notice that in 58, the answer was simply the positive way of saying what the original clue said negatively. That is, the original clue is M -> ~P and ~T (where ~ indicates negation). If M being in tosses P and T out, that means that L, S, and W are the only things that can be in with it (given that clue 1 already has said that M being in tosses H out). Phrasing something positively instead of negatively (or vice-versa) seems like one of the handful of standard things that LSAC will do to make a right answer with these.

[JasonR]

Good question that shit cost me a 178.

+1 for letting everyone know you got a 178... again...

RC Fail

[theZeigs]

One more question: what is the first PT that has this question type?

[LSAT Blog]

PT57 is the first to have this particular type of rule suspension question.

[willwash]

LSAC so far has made all of these questions relatively easy and there has never been more than 1 on a test (not counting my experimental which had 3 and clearly gave away that it was experimental).

Really? Good lord. I had an experimental games with only one (and yes, there's been precisely one scored version of this question in each test, including February, since it debuted last year).

I hadn't thought of the "use previous work to see if any of the situations you've drawn would be impossible with the answer choices" method. That's an interesting one that I'll have to try. I can't immediately find my notes for 57 and I still haven't gotten around to doing 59 yet, but I notice that in 58, the answer was simply the positive way of saying what the original clue said negatively. That is, the original clue is M -> ~P and ~T (where ~ indicates negation). If M being in tosses P and T out, that means that L, S, and W are the only things that can be in with it (given that clue 1 already has said that M being in tosses H out). Phrasing something positively instead of negatively (or vice-versa) seems like one of the handful of standard things that LSAC will do to make a right answer with these.

In that case, it would seem that the best thing to practice would be inverting each rule.

[bp shinners]

For these questions, they usually have you replacing a rule with two parts (H can't be first, but it must be before L; for example).

The correct answer will do 3 things:
1) Tell you the first part of the rule
2) Tell you the second part of the rule
3) Tell you nothing else.

If I remember them all correctly, 2 of the answers fail either part 1 or part 2 (they don't actually give you part of the rule), 1 of the answers gives you too much information (i.e. tells you F must be before H, when before we only knew it could be), and the last incorrect answer is split between either the first group or second group of wrong answers just mentioned. It's pretty easy to get rid of the 2-3 that don't give you everything you need out of the rule, and if there are 2 that give you too much, 1 is usually fairly obvious. After that, you can quickly test the 1-2 that are left against old answers.