Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Categories
- 33.4K All Categories
- 28.1K LSAT
- 17.1K General
- 5.2K Logical Reasoning
- 1.4K Reading Comprehension
- 1.7K Logic Games
- 70 Podcasts
- 191 Webinars
- 10 Scholarships
- 194 Test Center Reviews
- 2.2K Study Groups
- 112 Study Guides/Cheat Sheets
- 2.5K Specific LSAT Dates
- 31 November 2024 LSAT
- 18 October 2024 LSAT
- 9 September 2024 LSAT
- 38 August 2024 LSAT
- 28 June 2024 LSAT
- 4 April 2024 LSAT
- 11 February 2024 LSAT
- 22 January 2024 LSAT
- 38 November 2023 LSAT
- 43 October 2023 LSAT
- 14 September 2023 LSAT
- 38 August 2023 LSAT
- 27 June 2023 LSAT
- 30 Sage Advice
- 5K Not LSAT
- 4K Law School Admissions
- 13 Law School Explained
- 10 Forum Rules
- 641 Technical Problems
- 286 Off-topic
Complaint/Critique: Inconsistent LG approaches
brigitte
Free Trial Member
There are some situations where JY takes a different approach to a rule for no ostensible reason other than the fact that the ordinary approach would result in some difficulty that is not clear except based on hindsight. But this does not help someone facing the game for the first time!
The primary example of this is how we approach an ordering rule that has "or but not both." The usual way appears to be to split the game based on that rule, because it creates a "binary cut." This makes a lot of sense and is helpful. JY recommends this approach in PT83 Game 2, PT52 game 4, PT51 Game 2, and I might be missing some others. He occasionally solves without the split but also endorses the split and goes over it in a different video - PT78 Game 3. But what is troubling for me is that there doesn't seem to be an explanation of why one would approach this type of rule without a split rather than with the split. Obviously it's good to be able to do it both ways, but how does one know when one approach would be more effective than the other?
This problem reveals itself in PT61 Game 2, where JY does NOT do a split and does not talk about why he didn't do the split, even when one would quite naturally think about doing the split if one has been following the other videos. It turns out that the split, if done, is slightly messy and is not as easy to do as it normally is -- several students in the comments to the video note that the game was a lot harder with the split. But when asked why he didn't do the split, JY comments "Yeah, the P messes things up… It was a while ago, but I think that was why I didn’t link them up." This is definitely a good reason why the split doesn't end up being too effective, but it seems to be something that is only evident AFTER trying to split. We don't get to see the actual process of trying to do the split, seeing that it's not good, and then approaching the game without the split. Instead, it seems as if one should naturally know not to do the split. That seems like a hindsight based strategy rather than one made actionable for a student! I don't see any reason up front why we would not at least explore the split first.
Another example is the different approaches in PT73 Game 1 and PT53 Game 2. In PT73 we have 2 "or but not both" rules that create 4 possibilities. JY says it's a no brainer to sketch out those possibilities. Yet in PT53 Game 2, JY does NOT do the split and in fact mentions that he tried to the split but it wasn't helpful, so is showing how to do it without the split. But there doesn't seem to be any clear reason why a student approaching PT53 the first time after having reviewed the explanation for PT73 and similar games would proceed by not doing the split! PT53 presents 4 possibilities in almost the exact same way as PT73. It's like JY's showing the best way to do the game based on hindsight rather than showing a consistent approach that student could take to know up front which way is the best way to do the particular game!
Any thoughts?
Comments
Sorry this is a short answer to a very detailed question but... it pretty much comes down to foreseeing if the split is helpful. You normally do this by trying to envisage if the split let’s you make a reasonable number of further splits and/or useful inferences.
I’m on my phone so can’t link an example but imagine a sequencing game where you have a rule saying something like R or S (but not both) is in 7 (out of 8 slots). That info on its own is the kind that would let you split the board. But if splitting doesn’t let you make any more inferences, just having two game boards that look like this isn’t exactly helpful and so is a waste of time...
_ _ _ _ _ _ R _
_ _ _ _ _ _ S _
My thoughts are that there isn't always one better way to do a game. He shows both ways because some people simply prefer one method over the other. I'm a big advocate of splitting when you can and when it makes sense to do so. However, I think @LSATcantwin (Sage, scored a 171) said he likes to brute force games more often than split. (I think it was him? haha)
You have to figure out the best way that works for you!
[Admin edit: rule of thumb stated below]
A good thing to consider is how many different possible "worlds" and how many global questions there are for a particular game? If there is a simple game with 8 combinations, 5 questions, 2 of which are "If X is in ___ " it's probably better not to split. If there's a more complex game with 4 possible worlds, 7 questions, and 4 of them are just naked MBT/CBT questions, in that situation, it might be better to split.
This boils down to LG not being a science. It often hits me to do games differently the second, third, or fourth time through.
In the conditional sequencing game in PT 83,- JY decided not to split, but in an almost exactly same game in the past, he did decide to split (PT 73 game 1). JY is not a robot, so it doesn't surprise me that he'd do similar games, a few years apart, in different ways, It's vital to learn how to do both ways, anyway.
@"Alex Divine" Yep that was me. I set the game up, look to see if it is limited in any way, and then go right to questions. I then tackle each question individually. I can't think of a single time where I split the game board during a PT. I could see when a board could split, but still wouldn't do it for fear of wasting too much time looking for all the possibilities.
A good example actually was the December test: Game 4.
Based on a couple rules it looks like you could split the board. You actually could split it too, but it wasn't worth the time! It was much easier for me just to plug and play as it were. Based on that method I went -0...
@anonclsstudent If you have taken the December 2017 test you'll know which game I am talking about. It's the floor/cleaning game.
If not I can try and find one where I can see it would split, but wouldn't do it. I might need time on that though. I'm really swamped IRL and I have an interview with a law school tomorrow :O maybe I can cover this a bit in my AMA on Saturday!
Thank you for the feedback everyone!
I understand that it's useful to do both ways - split and not split. And I understand that there are factors to consider when it comes to whether a split would be useful.
But in the 2 problem games that I pointed out -- PT61 and PT 52 Game 2, my issue is that one of the approaches -- splitting -- is less effective, but one cannot tell that up front when deciding whether to do a split or not, and it would be entirely reasonable IMO for someone who has been following the curriculum to think a split makes sense. Turns out, it's not too helpful, but there's no explanation of what factors would lead one to know that at the start. It seems that JY's explanation chooses to forgo the split based on hindsight knowledge of the game. His own comments seem to indicate that (essentially when he says he tried the split first but it wasn't helpful so that's why he's showing the non-split version). That's the crux of my critique. I would have liked to see a discussion of what factors could lead a student up front to NOT prefer a split, or at what point when attempting the split that JY realizes it's not helpful and why -- this is because that's the reality a student would face when doing the game.
Sounds like a reasonable critique, but I definitely remember at least some videos where he discussed when it is a good idea to split. Sometimes it is hard to tell. If things immediately link up to your new information from the split, it is usually good. I aired on the side of splitting since if you write fast and small there isn't much of a cost to splitting and it makes some games a lot easier.
I have never felt like splitting stopped me from solving a game(only wasted a little time), but have definitely felt like it has saved me on a few.
I think you are right. And it's because they are his explanation videos. I think he covers what to do in real time (whether or not to split) in some Live Commentary videos.
Yes! I think also a lot of times, JY doesn't split because the rules aren't particularly restrictive. I wasn't able to see this when I first started, but as I've gone on it's become easier to see-- "ah, with these 3 rules, even if I do use Rule #1 to split it up into three different worlds, rules #2 and #3 won't add anything into whittling down these game boards." It's taken me some time to see, but I've gotten a lot better at being able to tell closer to the start when rules are more restrictive, and when making more worlds is beneficial, versus when the rules are quite broad and drawing out many game boards doesn't necessarily guarantee I'll be able to fill in anything.
@anonclsstudent I get exactly what you're saying. I often say, "how would I have known that?!" If I completed the game with no errors within time I don't worry about it. There's no concrete way to do the games. It's really just whatever works for you. I watch the videos and if I feel JY's way was better for me for that particular game I'll go with it, but if not, I stick with my way.
There are many different ways to skin a cat. LG is amenable to multiple methods. Be adaptable and seek to find thr methods that work for you. I know that I dont diagram games with a chart like JY does because I find it more time consuming and less intuitive than mine.
A sensible critique @anonclsstudent. And I promise you're not alone here. You're simply saying something out loud that lots of other students are wondering too.
I agree with your main point: strategies ought to be predictive, not retrospective. I try to hew to that principle, but don't always. A good example of a transgression is 38.G4 where I present a star as the game board, hehe. Retrospective. I only came up with that after a few retakes. I don't expect anyone to solve the game timed in that way. But, for that specific game, I really liked the elegance of the solution. So, I thought it would be worthwhile to share, even at the cost of violating the principle. I really should do another lesson that tries to solve the game with a more realistic game board, but I also have a ton of other videos that I need to do, so back of the line.
My general response is that in the older videos, I had not thought of many things. One of the things that I had not yet thought of was a rule of thumb for when to split. I just kind of felt it. That's why I didn't say anything in those videos, because I didn't know what to say. But these feelings are amenable to scrutiny. I tried to translate these intuitions into something explicit. That's when, after much time had passed, I hypothesized the rule for when to split. (See @"Alex Divine"'s comment above that states the rule. You're basically counting the number of SBGs against the number of "naked" questions.)
So, for the newer videos (which correlates with the newer PTs, though I do remake new videos for some of the older PTs) I do try to mention the rule of thumb for when to split, to explicitly demonstrate the predictive rule in action. I also regularly do two lessons now for when a game can be split, one to show how you'd approach it with a split and another without. (Can you see both lessons, e.g. 83.G2, with your free account?)
Re: 61.G2, if you apply the rule, it slightly favors splitting. There are 6 questions, 1 acceptable situation question and 1 substitution equivalence question. Generally, they don't count, because you do them as you do the rules. (This is another new technique that you won't see in the older videos.) We're down to 4 four questions. 2 questions give you additional premises, so we discount them by...something. Let's say 50%. 2 questions are naked / global / no additional premises. So we have a total of 3 questions. If we split, we'd have 2 SGBs. 2 is less than 3, so it weighs slightly in favor of splitting. At 4:00 in the main explanation, I tell you what the split would look like without drawing the split out because if drawn out, it would look awkward, with sequencing lines criss-crossing. And here I've implicitly called upon another principle of LG: visual clarity, though I should have explicitly mentioned this in the lesson. This principle says don't create a visual mess. So on balance, I don't draw out the split because you're already sort of looking at the split anyway. At 1:58 in the live commentary video (which free accounts cannot access), I mention this point again about being able to "see" the split boards without drawing them out.
Re: 53.G2, if you apply the rule, it does not favor splitting. There are 6 questions, 2 acceptable situation questions. We're down to 4 four questions, 2 of which give you additional premises. So again, 50% discount, we're looking at 3 questions. This time, splitting generates 4 SBGs. So slightly favoring not splitting.
Re: 73.G1, it's a close call. 7 questions, 1 acceptable situation question. Down to 6. 5 give us additional premises, so with a 50% discount, that's 3.5 total questions against 4 SBGs. Slightly favors not splitting but it's a closer call. If you did not split this game, I think you'd be fine precisely because 5 questions give you additional premises.
I'll also reiterate some of the other comments. Split v. not split is a judgment call, you do want to feel it. Even the rule of thumb, which is better than nothing, has judgment built in.
A separate point: if anyone has a different predictive rule for when to split v. not split, why are you not sharing? We'd all love to see something better than what I've come up with.
Finally, these "principles" or "habits" that I allude to can all be found here: https://7sage.com/lesson/logic-games-habits-for-speed-and-accuracy/
Knowing when to split and when not to split is an essential skill that, at best, may have general guidelines. Because there aren't hard set rules, any discussion of a game will be open to different interpretations on this point. Personally, I view splitting as simply a function of rule representation and little more. So for every rule, I simply consider the best representation: On the Master, in the rule list, or as a split. For most rules, the answer is clear. For many, it's a matter of context and experience. Context is the other variables in the game which lend weight to the decision, and experience is your ability to recognize and evaluate the context. With my experience, I'm usually able to make the correct call here, and I think a lot of what I'm able to do upfront would only be clear to most others with the benefit of hindsight. It's a skill that can be honed to a very high degree of ability. Sometimes I still make the wrong decision, but since it's just a function of rule representation, it's simply a matter of effectiveness rather than of correctness. I'll still get through the game successfully, I just could make things slightly easier on myself by having made a marginally better decision.
I have two primary considerations in deciding to split:
Will the other rules apply pressure on the split game boards?
Basically, will the split game boards create useful restrictions that will allow me to use the remaining rules to push out actionable information on the splits? If not, then I'm just trading an empty master game board for empty split boards. This isn't useful and is normally a waste of time and a bad representation.
How comfortable am I with the rule?
Sometimes, I'm just not feeling great about a rule and I'll split just to be rid of it. Maybe the most frequent example of this is when there is a single conditional rule on an otherwise straightforward grouping game. Experience has taught me that I do a relatively poor job of accounting for whether or not the rule triggers. So, if I'm worried that a rule is going to cause me problems, I'll split on it. This effectively eliminates the rule from consideration because it's accounted for. With a conditional, I simply trigger it on one split and fail it on the other. That's everything that could happen, so I use my splits from there and scratch the rule off of my list to remove the redundancy that I no longer have to account for. I love eliminating rules I don't like.
So that's splitting as a function of rule representation. For inferences, you have the exact same decision to make. They are simply unstated rules and they need to be represented the same way: on the master, in the list, or through a split. My same reasoning applies.
From here, there is often a decision on splitting which goes beyond rule representation. One is when I can see that I can completely solve out the game. Sometimes, you can see exactly how you could represent every world onto a number of sub game boards. Solving the entire game is not something I do lightly. I'll only do this in two situations:
The first is when I determine that the game is going to ultimately require it anyway. Game 4 from PT 1 is probably a good example. I just don't know what to do with this game when it's left unsolved. It's the type of game that I can tell will end up solved in the end because the questions will force me. How do I know? Context and experience. It's just that kind of game. If you're going to end up doing it anyway, it's better to do it up front.
The second is when I feel really shaky on the game. It's a damage control strategy that represents a failure that needs to be managed. A good example is the bus game from PT 36. Solving this game--splitting it into all possibilities--is not optimal. For me though, it's the only way I can get through it. Like I mentioned, I'm not in love with conditional rules on grouping games which makes this game something of a nightmare for me. Solving this game requires absolute mastery of splitting mechanics and fundamentals, and it's really easy to get lost. So if you've not directly practiced how to split cleanly and procedurally, this is quite dangerous. Try it. It's a good exercise. Anyway, this is my least favorite situation to use splitting. It's a crutch, and it should only be used in a damage control situation where you've failed to do things right and it is a more viable solution than other damage control strategies you may could use.
There's one other situation in which I will split that is neither a function of rule representation or solving. On particularly hard or weird games, my LG procedure has a contingency to deal with the anxiety/challenge. I walk through a few steps just to deal with the psychology of the situation: Breathe, remind myself I'm a badass at LG, that kind of thing. Then, I'll run a few scenarios. This is just working with the game board a bit which does several things: It familiarizes me with the game mechanics, it demystifies the game, and it provides me with a few boards to work reference later. This is huge to have. So to run these scenarios, I'll usually look for an opportunity to split the boards. The movie theater game from PT 70 is a good example. It turns out one of the game pieces is clearly restricted to three placements. So I split on that and then work with each game board a bit. This is a really tough game, and I find that this split really helps me.
Hope this is useful!
Does "splitting" mean to create an entirely new series of hash marks upon which your variables will rest?
Essentially, a "universe" diagram but one that is separate from the previous universes?
Example:
https://i.imgur.com/qr5Fk83.png
Are these "splits"?
Yup!
Okay, cool. I haven't seen all the LG videos, so please excuse my ignorance.
What about placing your variables below your game board?
Like this:
https://i.imgur.com/rWs2fbc.jpg
Does JY ever employ this approach or are all universes represented above board and as 'splits'?
I find this relieves a lot of the pressure of whether or not to split, because for me to split all I have to do is make a new row.
No, I think he usually does them on separate boards with the number line on the bottom. That's a neat way to do it though!
@goingfor99th Did you drill all of your games on scrap paper? If so, that's really neat. Did you ever find it to be an issue on the real test? I guess I'm asking if you felt at any disadvantage from having drilled games on scrap paper as opposed to on printed out games.
Also did you use 7Sage to prep or use something else? I forget?
I did a ton of journal work but I didn't start out doing that. When I compiled my "LG Journal," which those images are examples of, I was nearing the end of my LSAT preparation. That was probably at least the second or third time I had done those games, and I worked on that journal for a couple months, 3-4 months prior to my official June. In the two months prior to my test, I did all LG from PTs 38-70 on actual LSAT printouts as timed sections.
I believe the LG Journal was a key to my success. I would browse that thing all the time in the months leading up to my test. It allowed me to slow down and take everything in as a high resolution standalone game board.
I used Powerscore methods, worked through the Cambridge LSAT packets, had a tutor help me out from time to time, and supplemented with 7sage, Velocity LSAT, and a few other sources. I tried to adopt any approach I found helpful.
Not exactly sure which question @"Alex Divine" answered when he said "Yup!", but since I use both JY's method and your method, allow me to give my take on things. When it comes to the videos, splitting usually means sketching out an entirely new board on which to rest your variables. I view each split as a separate machine, clearly indicated by the visual space on the page. Your way of doing things was hard to wrap my head around at first, but I soon came to see that you took each different/slightly varying machine and neatly combined it into one board by marking the different sections of the machine (the different universes) with a new row. Therefore your boards are many, but at the same time they are one, in one sense it is a split and in another sense, it is not. I find your method to be more efficient to use in some cases, especially for simple sequencing games.
I hope my explanation makes sense to you. If it doesn't, we can share an awkward laugh about it later, haha.
Well put! :] I understand totally.
@JerryClarke242 I was saying "yup" to @goingfor99th 's question about whether or not what he was describing was splitting
Oh, ok. Thanks for clarifying.
I think different structures are fine as long as you know what you're doing and stay consistent. I wouldn't call this a split, although it seems fine as a hypothetical. Splits leave you with all worlds accounted for which this maybe only technically does. A split would take one variable, place it everywhere possible, and thus break the master into several complete possibilities based on the chosen variable. I don't think this is a good game to split, but maybe a split here would create three different game boards to account for every possible placement for F in 1 2 and 3 respectively. Nothing outside of those three game boards would be allowed to exist, and so the split would account for all possible worlds.
That game board is not exhausted. I think that game was more efficiently solved by focusing on the sequences rather than the game board itself.
Is this something more like what you're describing?
https://i.imgur.com/rFgNpag.jpg
This is another game I've mostly exhausted, but this one is grouping:
https://i.imgur.com/sXFqqS3.jpg
These are very helpful questions to ask to determine whether to split.
I use to do sequencing/liner games like this with the numbers on top and many rows below, and some times it works great and saves a lot of time. However, the problem with this approach is that building in and account for uncertainty (could be true elements) is much more difficult because there is no space to do so as a result of the compact design of the board. In many instances, your frames are not going to be all the way complete, so you will have to represent the uncertainty within each frame to properly symbolize the possibilities. 1 frame may actually represent multiple worlds in fact, and this is the mechanism by which the test derives could be true questions from the moving parts. I would recommend drawing each linear diagram out individually so you can represent the rules, variables, and other uncertainties/moving prats directly on, and around the board. You can create less frames this way too, since you don't have the have a diagram for every possibility, since it is built in.
Yeah, honestly can't say I've ever had that problem. (I'm not entirely sure what you're describing, though tbh.) When I need to, I leave room for internal diagramming. Very few games require internal diagramming, in my experience.
How many games have you done? This is a consistent feature in many of them, especially the more complex/harder ones.