770 (Q50, V45) < 3 weeks prep: Journal, Advice, Materials

Update: I got 5.5 on AWA and 8 on IR.

I just made an account to post this. I wrote my GMAT on Saturday (29 June) and got a score with which I’m pretty satisfied. As the title says, I put in a bit less than 3 weeks of actual studying, but nevertheless benefited from these forums during that time (what materials to buy, question explanations, AWA template, etc); hence, I decided to get out of lurker mode and contribute a debrief for posterity.

It’s pretty long so here’s a table of contents if you want to skip to certain parts
Table of Contents
1. Background
2. Materials
3. The Journey: Schedule and CAT Progress
4. Test Day: Game Plan and How It Went Down
5. Strategy Advice (AWA, IR, PS, DS, RC, CR, SC, timing, warming up, scratchpad use)
6. Efficiency Advice (getting the most out of your prep)
7. Advice on the Finer Nuances (Exercise, Diet, Music, etc)
8. Review of Materials (Pertains largely to the difficulty levels of various materials and my understanding of the algorithm used in MGMAT’s CATs)

1. Background
I finished my undergrad in business this semester and figured I would get my GMAT out of the way before I started working full-time in the fall. I actually want to go to law school (I’ve written my LSAT as well), but some factors pushed me to also lock in a GMAT score:

• it wouldn’t hurt to have one in case I want to do a JD/MBA or drop law altogether,
• given my business degree, a lot of my friends were doing the GMAT, and
• I’ll be working in consulting, so b-school is the bridge more often crossed

I was born in a foreign country and immigrated to North America at a young age. English is my 3rd language, but today it’s pretty much my mother tongue. My math was always quite strong in grade school, being 2-3 grade levels ahead. I took several math courses in university, but they were much more abstract and not relevant to the GMAT. In terms of GMAT math, I would definitely have been better at it back in 11th or 12th grade than I am now.

Initially, I was willing to put in one month of studying, but really wanted to keep the investment to a minimum (2-3 weeks). The very first thing I did was to Google “gmat in 2 weeks” and saw quite a few guides for such an endeavour, so I knew it was possible to get a good score in this timeframe. I also arranged some chats with people I knew who had gotten 99th percentile scores in a similar timeframe. I hope this debrief will help others out there who plan on tackling the GMAT with a short time frame for prep, but of course I hope it’s helpful to anyone who’s looking to score well .

2. Materials
Standard recommendation as per gmatclub (best-gmat-prep-books-recommendations-74310.html):

• Official Guide, 13th edition
• Manhattan GMAT Complete Strategy Guide Set, 5th Edition
• Manhattan GMAT 6 CATs
• GMAT Prep 2 CATs
• Manhattan GMAT Test Simulation Booklet w/ Marker
• Excel – I tracked everything very rigorously on excel. The data helped me be more efficient/targeted with my studying. I can probably spit out any stat you ask…for example, I did 1705 questions in total over the course of my prep, of which 198 were CR.

3. The Journey: Schedule and CAT Progress
High level overview of prep was basically 2 weeks to go through all the MGMAT guides while regularly doing CATs, at which point PS, RC, and CR were >90% accuracy. This was followed by several days of drilling DS and SC using the OG, and doing my final practice CAT 2 days before game day.

Since I didn’t prep for that long, I can actually provide a day-by-day journal of what I did. On any given day, I probably studied for 3-5 hours, broken into two sessions (ie. 1.5 hours in the afternoon and 2.5 hours in the evening). I’m very easily distracted/like to look for distractions, so typically studied while on Facebook . Total study time probably came out to around 80 hours.

This is also where I admit it wasn’t technically under 3 weeks if you count from the very beginning. I took my first CAT one month before the test day and another one a week after that, but didn’t actually start studying content until 20 days before the test.

Journal (CATs have been bolded):
29 May – MGMAT CAT 1: 740 (Q47, V45); -14 on quant, -9 on verbal. Was quite happy with this, but was afraid it might have been a fluke. Figured I’d take another CAT the week after. In retrospect, this 45V is inflated; you’ll understand when I talk about the algorithm in MGMAT’s CATs.
6 June – MGMAT CAT 2: 730 (Q45, V44); -16 on quant, -8 on verbal. Confirmed baseline wasn’t a fluke, used the assessment reports to figure out where I needed to improve, and went away for the weekend.
-----
9 June – This is the day I actually started studying, kind of. I skimmed through the RC guide during the car ride home. I didn’t find it useful. I put it away when I got home and didn’t touch it again after that.

10 June – Went through Number Properties guide. Very good content. Went 51/54 on the problem sets.

11 June – MGMAT CAT 3: 750 (Q48Q, V44); -12 on quant, -10 on verbal. Happy to see the quant score go up.

12-14 June – Went through Critical Reasoning and Algebra guides. This CR guide was a simplified version of Manhattan’s LSAT CR guide, good refresher. Again, algebra guide had great content. Went 46/49 on the CR problem sets and 95/98 on the algebra problem sets.

15 June – MGMAT CAT 4: 720 (Q48, V40); -13 on quant, -8 on verbal. I freaked out at this Verbal score, which led me to investigate MGMAT’s algorithms. Discussed under review of materials.

16-17 June – Went through Sentence Correction (Ch 1-10) and Geometry guides. For SC, most of the chapters were very useful. As a fluent speaker, I found chapters 7 (verb tense) and 9 (idioms) were a bit too technical/overwhelming, so I just skimmed these chapters. Went 60/62 on the geometry problem sets.

18 June – MGMAT CAT 5: 750 (Q48, V44); -11 on quant, -8 on verbal. I got 12/12 on this IR. I guess that was a good, but unnecessary accomplishment.

19 June – Finished off SC guide and started Word Problems guide. Went 167/184 on SC problem sets.

20 June – MGMAT CAT 6: 780 (Q51, V45); -4 on quant, -6 on verbal. This was an exciting result, but I realized it needed to be taken with a grain of salt given my understanding of MGMAT’s algorithms. Again, I discuss this in review of materials. Basically, I had exhausted MGMAT’s question bank of 700-800 level questions mid-way through this test.

21 June – Finished off Word Problems guide and went through FDP guide. Went 57/60 on word problems problems sets and 67/68 on FDP’s.

22 June – skimmed the MGMAT IR and Essay guide. Pretty useless. Did 2 RC passages from the OG and apparently nothing else according to my notes…

23 June – GMAT Prep CAT 1: 770 (Q50, V44); -7 on quant, -8 on verbal. So I didn’t get to review this CAT because my sister closed the program when I stepped away from the computer. I threw a fit and didn’t do anything productive that day.

24-26 June – Studying was getting boring. Main focus these few days was on drilling DS and SC questions from the OG. Did 100+ DS and 50+ SC questions during this period. I did MGMAT’s question banks for number properties and geometry as those were my weakest sections. I also did the last 10 questions in the OG for PS and CR, but found them too easy (and didn’t miss any), so I didn’t bother drilling those. In addition, I made sure to do 1-2 RC passages each day. They also seemed easy, but I wanted to make sure I didn’t go a day without doing RC this close to the test. I went 50/51 on these passages. On 25 June, I also did a “GMAT Prep CAT 1.5”. Since I didn’t get to review my previous CAT, I decided to do just a timed verbal section to see what kind of SC questions I was getting wrong. Was annoyed when I got repeat questions from the previous CAT (I’m assuming there’s a limited bank of hard questions). Finished the section in under an hour. Got a verbal score of 44, made 6 mistakes.

27 June – GMAT Prep CAT 2: 760 (Q50, V42); -6 on quant, -6 on verbal. At this point, I was REALLY bored of studying and just wanted to write the real thing. I rushed through this CAT and finished both sections with well over 10 minutes to spare. The verbal score was lower than I would have liked, but didn’t really bother me at this point. I did a genuine practice AWA essay (up until this point, I’d just been fooling around during the 30 AWA minutes). I gave myself 25 minutes and typed it up in notepad so that there was no spell check. I then copy pasted the essay into MS Word: word count was 618, Flesch-Kincaid level was 14.0, and there were a few typos.

28 June – Watched A LOT of YouTube. Re-watched the Hangover (first one, of course). Did a couple OG questions for each section throughout the day to keep the wheels spinning.

4. Test Day: Game Plan and How It Went Down
Game Plan – Given how my prep had gone, I had a pretty defined game plan for the day. I knew exactly what I was going to eat, what songs I would listen to, and exactly what would happen during my breaks. I would set up my scratchpad during the tutorial at the beginning, and get by the AWA and IR without exerting too much effort. My goal was to make 0 or 1 mistake on PS, RC, and CR, and hope for the best on DS and SC. Was aiming for 50+ on quant and 44+ on verbal for a 760+ score.

How It Went Down – My test was at noon. I had French toast and peanut butter at 10.30am. Cut up some apple slices and packed Gatorade. Did a couple questions from the OG to warm up. I arrived at the test centre 30 minutes early. I had 2 slices of apple, listened to my pre-selected pump up songs, and drank some water. It was a busy day for the test centre, so I started about 10 minutes late.

AWA – I got a straightforward prompt related to profitability. I threw down my template, picked 3 flaws, and free wrote for a while until I figured I’d passed 600 words. Proofread the essay, and had some time to spare, so I sat there and took a mini-break.

IR – my IR section was very straightforward. Definitely easier than MGMAT’s and a bit easier that the one that comes with GMAT Prep. I managed to solve every question with about 3 minutes to spare, so I took another mini-break before submitting my last question.

Break 1 – ate 2 slices of apple, went to the bathroom, drank 50mL of Gatorade, chilled in the waiting room, and went back in with 20 seconds left in my break.

Quant – the beginning was shaky here. It was taking me longer than it should have to solve the easy PS questions and DS felt rough. I was 5 minutes ahead of the clock after 10 questions so I took a little breather to calm myself down. Everything was smooth after that, but I had a weird feeling in the back of my mind I’d screwed myself a bit because I wasn’t getting any hard questions. I hypothesized that I’d made some dumb DS mistakes in those first 10 questions that dropped my estimate and I was fighting back up for the remainder of the test. I finally got some challenging questions for the last 3 questions. I knew I got them all right, so was hoping the recovery was enough to get me back up to 50 or 51, though I knew there was a chance I might have landed in the high 40s. I’d say PS went mostly according to plan – there was only one problem I couldn’t solve and had to guess on. Since I seemed to be climbing up the scale and didn’t get hard questions until the end, I finished the section with about 5 minutes to spare, so took another mini-break before submitting my last question.

Break 2 – ate 2 slices of apple, went to the bathroom, drank 50mL of Gatorade, chilled in the waiting room, and went back in with 40 seconds left in my break.

Verbal – the whole section went very much according to plan. I got 2 long RC passages, so had 14 RC questions in total. I’m fairly certain I got them all right. For CR, there was one question where I was split between two choices 60/40 and I’m fairly certain I got the rest of the questions right. For SC, I was finding the splits well and catching the tricky pronoun and modifier questions, but there were still a few I was uncertain about. I was doing 10 question blocks in 16-18 minutes, took a 1 minute break after question 20 just to calm myself down, and finished with about 5 minutes to spare.

Score – as I was doing the biographical section, I felt good about my verbal performance, and was a bit worried about whether I had recovered sufficiently to have a strong quant score. When I went to the score screen, I was relieved to see the 50Q and 770 total. Shortly afterward, I felt like I would have gotten a 51Q if I wasn’t shaky at the start, which would have given me 780. That said, the counterargument is that my verbal could have just as easily been a 44, and there really is no discernible difference above 760. I honestly prefer having scored 50Q/45V than a 51Q/44V because I really wanted the 99th percentile in verbal, so I’m satisfied with my results.

5. Strategy Advice (AWA, IR, PS, DS, RC, CR, SC, timing, and some other stuff)
AWA – it’s not worth putting much study time into prepping for the essay. I put in 4 hours of prep – the 30 mins during each of my 8 CATs. Basically what I did for the first few CATs was put together my template, then for the next few CATs, I practiced free writing my paragraphs. I wrote real essays for my last 2 CATs. When I realized I couldn’t retrieve my essay in GMAT Prep, I wrote my essay in notepad for my final CAT. For my template, I used the one posted here: how-to-get-6-0-awa-my-guide-64327.html, but made the sentence structures more complex and used bigger words so my intro and conclusion had a Flesch-Kincaid level over 13.5.
My strategy was to identify the conclusion of the prompt and pick 3 flaws really quick. Then I would dump my pre-written intro and conclusion while thinking about my body paragraphs. This would take about 5-7 minutes. Then I went and did free-writing for my body paragraphs, which entails me writing the topic sentence of the paragraph and then typing whatever came to mind for 5-6 minutes. So the whole writing process would take ~20 minutes and that would put be over 600 words. I purposely used very long modifiers and complex sentence structures to drive up my word count and Flesch-Kincaid level, as I believe those are the two main drivers of the essay mark.

IR – having gone through countless case studies during undergrad, I found IR to be very straightforward. Even if that’s not your background, I think the prep you put in for quant and verbal are sufficient for IR, so there is no need to prep IR separately. The IR sections you write during practice CATs should be more than enough to get used to the question types. For instance, after doing a couple IR sections, for a table question, you’ll get a knack for recognizing immediately which column you need to sort to get the required answer as quickly as possible. I would also say don’t over-rely on the calculator. For a lot of table and graphics questions, the answer choices are so far apart (ie. 10%, 20%, 40%, 55%, 70%) that there’s absolutely no need to use the calculator. Same thing with finding medians and means of two different data sets. Most of the time, it’s pretty obvious what the answer is without using the calculator.

PS – on quant, this is definitely the place to gain points. I think it’s very realistic to aim for 100% accuracy for PS. It’s a standardized test and every question is written to be solvable within 120 seconds, so the difficulty level has a ceiling. Learn the concepts that you need to know (MGMAT is sufficiently comprehensive), develop comfort with your toolkit, and be confident when you approach problems.

My process: read a question, know exactly what it’s asking for, go to scratch pad, and solve. Once I had an answer, I looked for it in the answer choices, double checked that I’d solved for the right thing, and moved on. If my answer wasn’t once of the answer choices, I’d double check that I solved for the right thing and double check my calculations. If knew right away what I needed to fix, I’d redo the question, but if not, then eliminate answer choices and go for an educated guess, but hopefully it doesn’t get to that.

DS – I found this to be tricky. It’s important to have a systematic way of approaching these questions – the MGMAT AD/BCE method was good, but sometimes it’s easier to tackle statement (2) before (1), so stay flexible. This is one question type where I think volume helps. The more questions you see, the more quickly you’ll know how to approach any given problem, like whether you should develop the question beforehand and how to pick hypotheticals. The hardest aspect was developing a way to think incredibly abstractly and flexibily so that you pick examples that are representative of a spectrum of cases (if you’re familiar with math proofs, think “without a loss of generality”). Realized that for DS problems, certain topics are emphasized more than others, such as odd/evens, positive/negative, geometry, inequalities, etc.

RC – this is the question type where prep will have the smallest impact. I found the GMAT RC questions to test more fundamental understanding and not as much inference as the LSAT. For instance, the GMAT has less emphasis on contrasting views and didn’t ask anything ridiculous like “if the passage had one paragraph added to the end, which one of the following would be the first sentence of that paragraph.” Given this, I think the important thing in RC is to invest the time upfront to UNDERSTAND the passage. Do everything you can to make sure you understand it. You should not skim ever. You should read every single word. Take notes to force yourself to read actively. If your eyes went over a sentence and you didn’t understand 100% what that sentence said, re-read it until you understand. If you truly understand a passage, you’ll answer every question in about 30 seconds, so it’s worth investing the time upfront.

CR – read the question stem first and then the prompt. Most important thing is to focus on the conclusion. After you read the prompt, if someone were to turn off your screen and ask you what the conclusion is, you should be able to restate the conclusion immediately. Eliminate wrong answer choices instead of looking for the right one – it’s must easier to do that at the higher difficulty levels. If the conclusion is clear in your mind, choices with no bearing on the argument are very apparent. You’ll likely be left with 2 answer choices. Compare each independently with the conclusion, not with each other. After doing that, one of the answer choices will seem very obviously right.

Many prep books say you should have an idea of what the right answer choice should be before reading the answer choices. At the higher difficulty levels, I disagree. Yes, this works for 500 and 600 levels questions because it’s so obvious. Doing so lets you get through early CR question in under a minute. However, having a preconceived answer before reading the choices will hurt you with more difficult questions in both accuracy and time. Since there are so many alternative explanations to an argument, there are too many possibilities for you to get stuck on a certain choice. Furthermore, sometimes reading the answer choices will make you realize that you’d overlooked an aspect in the prompt because you either missed a word, or you previously thought it was inconsequential. This is often seen with adjectives like crimes vs serious crimes, and time periods. When a question begins by saying “in the last x years,” that might seem like background info at first, but in fact time could turn out to be an important aspect of the argument. Hence, you should approach answer choices with an abstract idea of what you’re looking for and keeping an open mind. If it’s a weaken question, think “I need to look for something that might break a premise or assumption” instead of something specific like “what if not all student play guitar?” This will help you distinguish poor answer choices from relevant ones.

SC – Here, you make a lot of gains initially, but I think it requires quite an investment to master. Kind of like the English language – dead easy to learn, impossible to master. Once you’re extremely comfortable with the major topics (subject-verb, parallelism, pronouns, modifiers), you’ll have made most of your gains in accuracy. The rest of it is nuanced topics like idioms and a lot of subjective rules where you’re learning GMAT grammar, and not necessarily English grammar, so I didn’t feel it was worth it to burn time doing that. Even if you had 100% understanding on that, you’d probably still be prone to mistakes due to some questions having such poorly writing sentences that you interpret its meaning differently from the correct answer choice. Hence, I just made sure I was getting most of my RC and CR right, and just hoped for the best with SC.

Timing – for AWA I’d say aim to finish the essay under 25 minutes so you can give yourself a mini-break. Most of us probably type well over 60 wpm anyway, so we can write 500+ words very quickly. For IR, you get an average of 2:30 per question. I’d say most questions need less than 2 minutes, and there’s usually one or two questions where you need to spend up to 4 minutes. Always remember that there will be a multi-part question, where timing needs to be treated like RC – it’ll take a while to absorb everything first, but the 2nd and 3rd questions can probably be answered within a minute. In net, multi-part questions gain you time as you’ll average less than 2:30 per question. Hence, you have to keep this in mind when you’re checking time. If you get a multi-part at the beginning, you should be ahead of the clock the entire section. If you’re a bit behind after question 8, but you still haven’t done your multi-part question, you’ll probably be fine.

For quant, remember that your first set of questions is going to be at a difficulty level below your actual ability, so you can gain time here, provided you’re maintaining accuracy. I always built up a safety “bank” of time and almost never put myself behind the clock. If you do the first 10 questions in 15 minutes, you’ll have a 5 minute “bank”. Knowing this, when you stumble across a challenging problem where you need to spend 3 minutes on it, you can afford to do so comfortably. Every time this happens, check your time to see how much bank you have left. When you feel your bank diminishing, start pushing your pace to build it back up. You should never feel the need to completely “throw” a question just to catch up on time. You should have a small bank left as you enter the last 5 questions (1 or 2 mins), so figure out at that point how much time you can invest for the end.

For verbal, I guess this depends on fluency as that impacts reading speed. I think most fluent speakers don’t have too much of an issue with timing. I try to use up the time limits so actually took it a bit slower than I could have, and paced around 16-18 minutes per 10 questions. Same thing with quant though, the early questions will gain you time, so you can build a time bank earlier on. This way, if you get a convoluted RC passage or a tricky CR problem where you need to re-read the prompt and sit there and think about it, you know you can afford to spend to the time you need to in order to get the questions right.

Warming up – always do warm up questions before a CAT to get the brain clicking. For me this was 3-5 DS questions before a practice CAT. On test day, I warmed up with 2 PS, 3 DS, 2 CR, and 3 SC. No need for excessive warm-up, since the IR section is pretty much a warm-up to quant, but still do something.

Scratchpad use – know how you’re going to use this guy ahead of time. Practise and refine your setup throughout your CATs. I set up my scratch pad at the very beginning during the 4 minutes you get to do your tutorial. My final configuration was as such.
Page 1: essay, first few words of every sentence in my intro and conclusion.
Page 2: IR, open space.
Page 3 – 7: quant, 8 questions per page for first 4 pages, then last 5 questions on last page. On first 4 pages, I left the bottom right empty as extra space as needed.
Page 8: verbal. ABCDE writing across the top, and then question number writing down in a column as blocks of 10. Right side left empty to take notes for RC.


6. Efficiency Advice (getting the most out of your prep)
Think marginal benefits – This is where MGMAT’s assessment reports and keeping an excel spreadsheet pay dividends. Decide on some performance metrics. I used accuracy as a proxy for my progress. I know it’s an adaptive test, but all else equal, you score higher if you get more questions right. I didn’t really see a better metric.


Study smart and efficiently. Try to gain something out of every hour you invest. Think of your time as a resource and how you can maximize your marginal gains from each additional hour of study. Apply this concept from macro to micro. If your splits are 40Q/35V, you’ll probably have an easier time going to 40Q/40V, as opposed to 45Q/35V, so study verbal. If your splits are 45Q/45V, then your gains are going to be made in quant.

Then break the sections down. At the beginning, my PS accuracy was ~70% and DS ~40%. So it was obvious my DS needed work. In terms of topics, you’ll see I didn’t go through the MGMAT quant guides in order. I did number properties first before that was my worst area conceptually. I then did algebra because that was my slowest area.

When I was using the OG, I didn’t do PS and CR questions because there was no point. For every PS/CR question I did, I could have been doing an extra DS or SC problem, perhaps learning new things, which wouldn’t happen from the OG PS and CR problems.

Reviewing CATs – You should review CATs the same day when the exam is still fresh in your mind. Don’t only review questions you got wrong. It is just as important to review question you got right. You need to understand why you are getting things right so you can replicate that process. You can understand why you got something wrong and avoid that mistake, but avoiding one mistake doesn’t preclude you from making another type of mistake. During practice tests, you should be flagging every question you aren’t 100% sure on, this way you’ll review questions you got right through an imperfect process or through guessing. Basically, go through the entire CAT from beginning to end. When you get to a question you got wrong, redo the question first before checking what the answer is.

Error log and reviewing older exams? – Ehhh. I kept an error log because it seemed like a thing to do, but I wouldn’t do it again. I don’t really think there’s a point in looking backward too much. Honestly, there’s 2 reasons you make mistakes – you didn’t know the concept so didn’t know how to approach it, or you made a careless mistake. There’s no need to overcomplicate this. There’s no point to reviewing RC and CR that you did weeks ago. Those questions are too “in-the-moment.” Old PS, DS, and SC problems you can probably review once down the road to refresh yourself on some concepts you didn’t know before (but surely you know them now ).

Don’t overextend – if you’re tired, stop studying. Studying with a tired mind is inefficient. You’re slower, you’re not thinking well, and you’re not absorbing content at your potential. If you’re getting tired, take a break or resume the next day. In school, I always thought pulling all nighters for academics was dumb. You’re better off just going to sleep and resuming in the morning. Your mind can probably do more in 1 hour well rested than you can try to force it to do in 3 hours when it’s groggy.

7. Advice on the Finer Nuances (Exercise, Diet, Music, etc)
It’s important to take care of your body while studying for any standardized test. I ran 6 days a week and ate relatively healthy (I usually eat whatever I want, resulting in a lot of junk). Fatty foods slow down the brain, so I took a bit of a hiatus from junk foods. I also cut out drinking almost completely during the study period, which was a drastic change from the final year of university. Since I wanted to prime the brain, healthy foods meant a more protein heavy diet, with a lot of fish.

For test day, Gatorade is often recommended because it helps replenish sugar levels, which is important for the brain. I also read somewhere that its effect on blood pH also reduces your need to urinate. But I’m not a science student so you should probably do your own research. My other source for energy/sugars during the test was my apple. I think any fruit will do the trick. Again, do you own research on this matter. I think diet is an important aspect of testing. I honestly never felt that tired during my testing as I was always replenishing my energy.

You also need some hobbies/activities/whatever to moderate your mental mood. I usually study with music in the background. Whenever I felt frustrated or needed to calm down, I would go for a run or play the piano. I also pre-selected my music for test day to pump myself up. I think it’s important for everyone to figure out what kinds of activities have what effect on their disposition and use that knowledge to keep your studying on track.

8. Review of Materials
Official Guide – use this as a question bank for drilling. If you’re aiming for 730+, the questions here are on the easy side. That said, the way you compensate for difficulty is accuracy and time. If you’re a 48+ on quant, you should probably be getting 95%+ accuracy, averaging less than 90 seconds per question on the OG PS questions.

They say questions are listed in order of difficulty, but I had a weird experience with DS. Most questions up to 100, I could do in my head and get right. From questions 101 to 141, I got 12 wrong. Then questions 142 to 174, I got all right. With RC, I think the OG equates “difficult” with “science passage.” I guess since they’re testing for fundamental understanding, science passage might be harder to grasp, but it’s still silly to put 6 sciences passages in a row at the end

Nonetheless, the OG is an extremely valuable tool since they’re official questions. It was particularly useful for SC, reading the official explanation. For something as controversial as grammar, I would not do any non-OG SC questions outside of CATs.

MGMAT Guides and Question Banks – MGMAT’s quant guides are very comprehensive. I don’t think I did a single problem where the concept needed to solve the problem wasn’t covered by Manhattan. SC guide was great, CR guide seems okay, and RC guide was useless (as is any RC guide). The difficulty level of the problem sets is low, but I think straightforward problems are good for solidifying concepts you just read through. Question bank difficulty is also low, with very few 700-800 level problems. I did the number properties and geometry QBs and stopped after that. Average time for questions was 70-80 seconds.

MGMAT CATs – I think everyone knows now that Manhattan’s quant sections are much more difficult than the real thing, and this is great. You’ve got to do questions that push you beyond the real thing. It’s also great because if you manage to control your timing on Manhattan’s CATs, you’ll probably be ahead of the clock on the real thing. For verbal, they had some sketchy/questionable CR and SC questions, but they were rare (one that pops in mind is they had a sufficient assumption question, which is something you see in the LSATs. But the GMAT only tests necessary assumption, so the question was asking for a necessary assumption, but the credited answer choice was a sufficient assumption.) They made their RC longer by making LONG passages. One CAT that I was actually pressed for time at the end, I got 3 RC passages that were each over 70 lines long, and accompanied by only 3 questions each…you wouldn’t see that in the real test.

My criticism of MGMAT CATs is their algorithm, which wasn’t too difficult to figure out. It’s based on WHEN you get questions wrong, and not really on question difficulty. Basically, they run a percentile estimate that adjusts as you get questions wrong or right. Then they give the 0-60 score that corresponds to the percentile you end the test at.

The issue with this is that their CATs overemphasize the last questions of a test. Figuring this out probably hurt me for the real thing as I should have been more careful with the first 10 questions, but never got in that mindset because the first 10 don’t have as profound an impact in a MGMAT CAT. As mentioned earlier, I figured this out when I got a 40V on my CAT 4, despite making fewer mistakes than when I got 45V.

When I got 45V, I didn’t make any mistakes in the last 6 questions, so my percentile estimate stayed at 99%ile. (And this was probably an inflated verbal score). Each time I got 44V, I got either #38 or #39 wrong. I would be holding 99%ile estimate for a string of questions up to that point, the mistake would drop me to 90% or 91%, then getting the last few questions right would push be back up to 97%, and with no questions remaining, that’s where I’d end. When I got the 40V, I had a 99%ile estimate from question 27 to 35, then made 3 mistakes in the last 6 questions, which left me at 90%ile. I doubled check this methodology with quant, and it was consistent. When I posted this on their forums, no one from Manhattan replied to my thread .

Finally, the MGMAT question banks are just a bit short of having enough 700-800 level questions. This hypothesis arose during my CAT 6 my quant section, when I did all 15 of my DS questions within the first half of the test. Since I was getting 700-800 level questions since CAT 1, I assumed that by CAT 6, the question bank had run out of 700-800 level PS problems, so it could only give me DS problems. I checked the difficulty for the questions for the latter half of my quant section, and they were all 600-700 level questions. Same with verbal – as I was doing 700-800 level questions since CAT 1, the latter half of my CAT 6 verbal section were all 600-700 level questions. And this is why my accuracy was so high on my CAT 6 and I got 51Q/45V. But since I understood what happened, I knew this was an inflated score.

Nonetheless, these CATs were AWESOME prep and I would strongly recommend these 6 to anyone (and already have). I’m sure it’s impossible to replicate the true algorithm used by GMAC and every prep company’s CATs have their own issues.

GMAT Prep CATs – these seem like the most accurate predictor, but have a limited question bank of high difficulty questions. So if you’re a high scorer, you really only get 2 shots at these. If you reset the question bank and do a 3rd exam, you’re going to see a lot of repeat questions. If you click “end exam” when you’re reviewing, you’ll lose all your data. Instead, click “end review” to be able to access the exam again at a later time. The practice 90 questions they give with the software are wayyy too easy. I did the DS and SC ones and average under 60 seconds per question.

So, that’s it! I’ve really enjoyed reading the other debriefs, so will probably be trolling these forums for the next little while. Happy to answer any questions anyone has.