GMAT 655, from 575 to 655, with e-GMAT

Question

I spent two months trying to learn my way out. What I actually needed was permission to lose questions on purpose.   
   The gap I could not explain   
  My first mock came back at 575 (V75, Q87, DI74), and the number was not the hard part. My untimed accuracy was fine. I would work hard questions on Neuron, which is e-GMAT's bank of official-style questions with full worked solutions, get most of them right, then sit a fully adaptive test and watch the same skills come apart. So I assumed conceptual issues: number properties, vocabulary, or just needing another five hundred questions. I never comparison-shopped a course. I had a decent base that hid where I was stuck. What I needed was something to tell me what to fix. More syllabus to cover was never the shortage.  
   Verbal was the two-month blocker   
  My worst habit had a name once I could see it: storytelling. I read fast, and when an argument left a gap my brain invented scenarios to make a wrong choice work. Test writers know intelligent test takers love building narratives, and they design choices that depend on exactly those unstated scenarios.  
  What broke it was an e-GMAT verbal lesson that reframed the section as a closed logical system: an assumption or weakener has to connect what is stated to the author's exact conclusion in one direct step. I drilled it on Scholaranium, e-GMAT's practice bank with topic-level analytics and timed quizzes, until pre-thinking replaced the instinct to narrate. Pre-thinking here means working out what the argument needed before I looked at the options. When two choices still looked identical I negated one mentally; if that did not destroy the author's bridge, it was gone in five seconds. My verbal sectionals ran 78, 80, 82, 83, then 84.  
   Burning down my error log   
  My mentor pushed me to keep an error log and I built one, but for weeks it was an administrative spreadsheet and nothing more. I logged the source, noted why C was right and B was wrong, and moved on. A completely false sense of security, because two days later the same logical shape arrived in a different question and I fell into the identical trap.  
  So I burned it down. I capped my volume on Neuron and Scholaranium at eight hard questions and gave each 15 to 20 minutes: stripping the decorative language, writing out premises and conclusion, then recording why the wrong choice had appealed to me. Familiar wording lifted from the passage, or a moral judgement I had quietly agreed with? I finished with 60 to 70 deeply analysed entries, each one a rule I reread before every mock.  
  The most useful number I saw came out of that same log. Across my RC medium cementing quizzes, which are the short exam-level checks e-GMAT sets after each topic with a threshold you clear, I averaged 65 percent at relaxed timing and 52 at standard timing. Identical difficulty, identical pool. Thirteen points, and the clock was the only variable that moved. That ended the conceptual-gap story for good.  
   My closed section was still leaking   
  I am a software engineer, and engineers are trained to solve every problem in front of us no matter how long it takes. On an adaptive exam that is lethal, because the algorithm is built to test resource allocation under strict constraints. On the 16 August mock my Quant fell to 84, in the section that supposedly needed no work, because I spent three and a half minutes on one problem mid-section and then rushed four closing questions I could easily have solved.  
  The fix came from two directions. I worked the e-GMAT course Quant modules in sequence, no jumping around: Number Properties, Word Problems, Algebra, Advanced Topics. Scholaranium's topic analytics told me which sub-areas were actually bleeding, so I stopped guessing. Then the Quant sectional mocks did the rest, because knowing a method is necessary and nowhere near sufficient. Timed execution comes from 45 minutes of a full section under a clock, repeated. My next Quant sectional came in at 89, and Quant read 90 on my final mock.  
   Data Insights, the section I had written off   
  DI started at 74, and what fixed it was coverage across every question type. I had been doing more of the one I found comfortable. e-GMAT builds Data Sufficiency, Two-Part Analysis, Multi-Source Reasoning and Graphics and Table Analysis separately, and the variety is the point: TPA holds up when you know a TPA question is coming and slips when it arrives mixed in under a clock. Cementing quizzes closed the concepts. The DI sectional mocks taught me to switch formats cold. My DI sectionals ran 76, 71, 75, then 82, and DI moved from 74 to 83. Hard MSR and hard TPA never fully came right.  
   The two weeks I thought were not enough   
  Thirteen days out I gave a mock, got 635, and emailed my mentor to say two weeks would not be enough. I was angry, mostly at myself, having been stuck at 615, 625, 635 for months.  
  What came back was arithmetic. My four mocks had summed 236, 245, 246 and 245 across the three sections, while my best-ever section scores at that point, Quant 89, Verbal 80, DI 83, summed to 252 against a target needing 250. Every number I needed had already happened on my account. Just never in the same sitting. What stood between me and the score was not learning.  
  That is what the Last Mile Push is for. LMP is closing-phase mentorship: a mentor holding your data and catching the behavioural patterns that are capping your score. No new content gets assigned. Mine was Jagadish, and two things changed the result. First, a strict one-minute-thirty circuit breaker, after which I stopped rereading the prompt, eliminated whatever was mathematically out of bounds, made a calculated guess, flagged it and moved on. Second, permission to miss two or three questions a section on purpose, provided the misses were deliberate control sacrifices and not careless blunders. For an engineer, that permission was the whole unlock. It preserved the cognitive bandwidth to execute from the first question to the last. If you are in your closing weeks, look into LMP.  
   Mocks bounce   
  575, 635, 645, 635, 695. The 695 was the 98th percentile, given on a browser that froze mid-section. Test day landed at 655 (V80, Q87, DI80). Short of the 685 I wanted, but 80 points above where I started, earned while coding nine hours a day.  
   What I would take from this

Good untimed accuracy plus poor timed scores is not a content problem. Stop buying content.  
  An error log written in five minutes is a diary. Give one question twenty minutes and ask why the wrong answer appealed to you. 
 
  Deliberate misses are a strategy. Careless ones are not.

Rashie1 · Answer

The last three lines from the debrief is very helpful! Congratulations on your score Shivam!

pintukr · Answer

Thanks for the insights into your preparations. All the Best.

	Target Test Prep - Save up to $599 $400 Discount + $99 Bonus GMAT Club Tests
	e-GMAT: 60% Off on GMAT Focus Get to 735+ from any starting point
	Magoosh increases your score, guaranteed - 15% off ALL Magoosh GMAT plans with code GMATCLUB15 + Free GMAT Club Tests.
	Manhattan Prep: $99 Bonus Score Higher on Your GMAT Exam. Guaranteed.
	Experts' Global - ~~$1720~~ $960 End-to-End GMAT Prep + App Support: 150+ Videos, 2000+ Questions, 15 FLTs +
	GMATWhiz - Save up to 44%! Plus get a $99 bonus from GMAT Club when you buy a course. Sale ends Sept 7.
	GMAT Club Tests - Free Included with every course purchase of $149+ View all deals

GMAT 655, from 575 to 655, with e-GMAT

Prep Toolkit

Top 5 GMAT Debrief Videos

615 to 715 in 15 Days (Ria's Strategies)

More than Quant/Verbal Abilities: Speed vs. Accuracy

745 with Only Self-Prep and GMAT Club

From 555 to 765: 210-point Improvement

Perfect 805 Score Debrief by Julia

My Rewards