Last visit was: 18 May 2026, 16:30 It is currently 18 May 2026, 16:30
Close
GMAT Club Daily Prep
Thank you for using the timer - this advanced tool can estimate your performance and suggest more practice questions. We have subscribed you to Daily Prep Questions via email.

Customized
for You

we will pick new questions that match your level based on your Timer History

Track
Your Progress

every week, we’ll send you an estimated GMAT score based on your performance

Practice
Pays

we will pick new questions that match your level based on your Timer History
Not interested in getting valuable practice questions and articles delivered to your email? No problem, unsubscribe here.
Close
Request Expert Reply
Confirm Cancel
User avatar
iPen
Joined: 08 Jun 2015
Last visit: 20 Dec 2020
Posts: 80
Own Kudos:
Given Kudos: 40
Posts: 80
Kudos: 109
Kudos
Add Kudos
Bookmarks
Bookmark this Post
avatar
Porkbun123
Joined: 28 Jul 2014
Last visit: 23 Dec 2018
Posts: 69
Own Kudos:
24
 [1]
Given Kudos: 18
Location: United States
Posts: 69
Kudos: 24
 [1]
1
Kudos
Add Kudos
Bookmarks
Bookmark this Post
User avatar
iPen
Joined: 08 Jun 2015
Last visit: 20 Dec 2020
Posts: 80
Own Kudos:
Given Kudos: 40
Posts: 80
Kudos: 109
Kudos
Add Kudos
Bookmarks
Bookmark this Post
avatar
Porkbun123
Joined: 28 Jul 2014
Last visit: 23 Dec 2018
Posts: 69
Own Kudos:
Given Kudos: 18
Location: United States
Posts: 69
Kudos: 24
Kudos
Add Kudos
Bookmarks
Bookmark this Post
iPen
Porkbun123
youre off base here dude. If you get the easy questions wrong the gmat isn't going to throw medium level questions at you at all. It will keep throwing easy ones until you get them right consistently.

But what if the experimental questions are medium/difficult, and you get those few right? It's not consistent enough to prompt more medium/hard questions?

And, what if you're prompted with a random medium/hard question to start off? Or, is it always an easy level question at the beginning?

You do know the purpose of the experimental questions? It's to help the testmaker determine the appropriate level of difficulty. Which means that those experimental questions you see arent "assigned" a difficulty at all because they haven't been determined yet.

And you most likely aren't going to be prompted an easy questions at first. You're prompted a medium level question. If you get that wrong you'll see a downward shift. If you get it right you'll see an upward shift. The Gmat is never going to test you at a high level if it doesnt think you belong at a high level. You have to "earn" your way there. If you think about the purpose of the adaptive cat, its designed to probe your true testtaking level on a precise scale. Throwing a difficult question at a novice isnt going to test for much. If the novice gets it right, it is probably luck. If the novice gets it wrong, then that was expected and now the test just wasted 1 question confirming what it already knew.
User avatar
iPen
Joined: 08 Jun 2015
Last visit: 20 Dec 2020
Posts: 80
Own Kudos:
Given Kudos: 40
Posts: 80
Kudos: 109
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Porkbun123
iPen
Porkbun123
youre off base here dude. If you get the easy questions wrong the gmat isn't going to throw medium level questions at you at all. It will keep throwing easy ones until you get them right consistently.

But what if the experimental questions are medium/difficult, and you get those few right? It's not consistent enough to prompt more medium/hard questions?

And, what if you're prompted with a random medium/hard question to start off? Or, is it always an easy level question at the beginning?

You do know the purpose of the experimental questions? It's to help the testmaker determine the appropriate level of difficulty. Which means that those experimental questions you see arent "assigned" a difficulty at all because they haven't been determined yet.

And you most likely aren't going to be prompted an easy questions at first. You're prompted a medium level question. If you get that wrong you'll see a downward shift. If you get it right you'll see an upward shift. The Gmat is never going to test you at a high level if it doesnt think you belong at a high level. You have to "earn" your way there. If you think about the purpose of the adaptive cat, its designed to probe your true testtaking level on a precise scale. Throwing a difficult question at a novice isnt going to test for much. If the novice gets it right, it is probably luck. If the novice gets it wrong, then that was expected and now the test just wasted 1 question confirming what it already knew.

ok, I see. I edited my last response since but it looks like you answered my question. Thanks.

Though, maybe it's still a good idea that GMAC offers a hard-level-only test in the interest of a shorter exam.
avatar
Porkbun123
Joined: 28 Jul 2014
Last visit: 23 Dec 2018
Posts: 69
Own Kudos:
Given Kudos: 18
Location: United States
Posts: 69
Kudos: 24
Kudos
Add Kudos
Bookmarks
Bookmark this Post
I think the best way for the GMAC to cut down on the length of the exam is to get rid of the useless IR and AWA sections hah. But if they're never going to do that I doubt they're going to touch the core quant and verbal sections.
User avatar
IanStewart
User avatar
GMAT Tutor
Joined: 24 Jun 2008
Last visit: 17 May 2026
Posts: 4,146
Own Kudos:
11,323
 [1]
Given Kudos: 98
Location: United States
GMAT 1: 780 Q51 V47
Expert
Expert reply
GMAT 1: 780 Q51 V47
Posts: 4,146
Kudos: 11,323
 [1]
1
Kudos
Add Kudos
Bookmarks
Bookmark this Post
Most of the explanations of the scoring algorithm that you'll read are not correct - they're dramatic oversimplifications of how the algorithm actually works. In many books, you'll find a 'tree diagram', and the '20 questions' analogy you post above is a variation on that diagram. Neither description is accurate. I dislike those kinds of explanations, because they make it seem that the GMAT is based on some simplistic "if/then algorithm" that someone just made up. And when people think the algorithm works that way, naturally they think there might be ways to "beat the algorithm" by figuring out how it works. None of that is true.

I've been meaning to write up a careful and precise description of how the algorithm works, but it would take me more time than I have right now. But in brief, GMAT scoring is based on a field of academic research that has been in existence for more than fifty years. It has a rigorous mathematical basis in probability theory. For each question on the test, the algorithm knows (from data collected when the question was used as an experimental question) the probability test-takers at each level will give a correct answer. Essentially, for each question, the algorithm has a graph, which maps ability against the probability of a right answer (if you want to see what these graphs look like, you can google 'Item Characteristic Curve' and look at any of the images). So when you finish the test, the algorithm knows what questions you got right and wrong, and knows the probabilities test takers at each level would have answered those questions correctly and incorrectly, and from there, using some heavy math, it works out your most probable ability level.

To answer your original question, the probability an 800-level will answer a 200-level question correctly will be extremely high - probably something like 99.9%. So it's very unlikely an 800-level test taker will answer one easy question incorrectly, and astronomically unlikely he or she will answer two or more of them incorrectly. The algorithm would simply never think that a test taker answering two 200-level questions incorrectly is an 800-level test taker, because 800-level test takers essentially never do that.
User avatar
iPen
Joined: 08 Jun 2015
Last visit: 20 Dec 2020
Posts: 80
Own Kudos:
Given Kudos: 40
Posts: 80
Kudos: 109
Kudos
Add Kudos
Bookmarks
Bookmark this Post
IanStewart
Most of the explanations of the scoring algorithm that you'll read are not correct - they're dramatic oversimplifications of how the algorithm actually works. In many books, you'll find a 'tree diagram', and the '20 questions' analogy you post above is a variation on that diagram. Neither description is accurate. I dislike those kinds of explanations, because they make it seem that the GMAT is based on some simplistic "if/then algorithm" that someone just made up. And when people think the algorithm works that way, naturally they think there might be ways to "beat the algorithm" by figuring out how it works. None of that is true.

I've been meaning to write up a careful and precise description of how the algorithm works, but it would take me more time than I have right now. But in brief, GMAT scoring is based on a field of academic research that has been in existence for more than fifty years. It has a rigorous mathematical basis in probability theory. For each question on the test, the algorithm knows (from data collected when the question was used as an experimental question) the probability test-takers at each level will give a correct answer. Essentially, for each question, the algorithm has a graph, which maps ability against the probability of a right answer (if you want to see what these graphs look like, you can google 'Item Characteristic Curve' and look at any of the images). So when you finish the test, the algorithm knows what questions you got right and wrong, and knows the probabilities test takers at each level would have answered those questions correctly and incorrectly, and from there, using some heavy math, it works out your most probable ability level.

To answer your original question, the probability an 800-level will answer a 200-level question correctly will be extremely high - probably something like 99.9%. So it's very unlikely an 800-level test taker will answer one easy question incorrectly, and astronomically unlikely he or she will answer two or more of them incorrectly. The algorithm would simply never think that a test taker answering two 200-level questions incorrectly is an 800-level test taker, because 800-level test takers essentially never do that.

I see, thanks for the info.

Hypothetically speaking, then, if I get all of the hard questions right, I won't be stuck with just really hard questions? So, the GMAT will throw in easy/medium questions for everyone and essentially there's a cap on the number of hard questions given to all test takers, since GMAT's algorithm can sufficiently determine your difficulty level based on a minimum set of hard questions. Otherwise, an exam full of 700+ level questions is going to be mentally taxing, to say the least...
Moderators:
201 posts
General GMAT Forum Moderator
474 posts