A question for the Maths buffs

This was a reply I got on Reddit .

The probability to guess right on multiple questions when you have 5-6 alternative answers is very very low. That means that you probably won’t inflate your score very much by guessing.

I’ll readily admit I’m no maths genius. I’ve come to the conclusion there would be a 20/16.67% chance of a random guess being right. On a 30 question test that would equate to 6/5 answers right.

Or am I going wrong with this?

1 Like

I dont know i was and still am super poor in maths… i hate freaking maths… equations are supet complex…!!!

Depends at how many questions there are, if just 1 question with 5 answers but the more questions there are the odds get worse lol

Found this https://en.commtap.org/mcprobcalc

Depends on the probability distribution, I guess. I don’t know how tests are developed or if teachers can pick and choose certain answers. When I was in high school, they told us to guess C if we didn’t know anything. Like it would increase our chances of success. Some answer choices are less likely to be picked. Hence, you don’t see a lot of E answers unless some fluke or something. If it’s a uniform distribution (all answer choices likely), it’s easy to calculate. If there’s 6 choices, you do 1/6 and multiply by the number of questions. If there’s 30 questions and 6 choices, the expected value is 5 correct choices (mean). I guess someone smarter than me will tell me I’m wrong.

When I went to university for math (after I transferred), we had partial credit, no multiple-choice tests. Kind of makes you a big boy lol. A lot of the tests averaged to 50% then it was curved. So a 50% was a B+/B.

If they want kids to learn they need to get rid of multiple-choice tests. I’m starting to think Common Core for math is good because it makes people think. I used to be against it, but now I learn towards it.

I had a trigonometry class when I tried community college years back I was dumbfounded

On that fateful day I took the university exam so I could get in the college…I studied a couple of days before the exam…in science they said time five minutes…I looked down and I hadn’t answered almost all the questions…I guessed on the remaining questions just to see if I could get some right and when I saw my average I made a 17 so I passed the exam but I would have flunked but science was almost an A…score…which brought my average up…true story.

I love multiple guess tests. There are usually 4 choices. You can eliminate two right off the bat. Then it becomes 50% that you guess the right answer.

Found this .

Another justification given for not using objective testing is that candidates can gain inflated scores due to guessing. If you consider the example above then a candidate would have a 1 in 5 chance of correctly guessing the answer. This suggests that it would be easier to pass an objective test.

However, the effects of guessing can be eliminated through a combination of question design and scoring techniques. Those that advocate objective testing suggest a number of strategies. For example, questions could be scored as normal, and a formula for “guess correction” could be applied at the end. There has been a lot of research in this area and a “standard” formula has emerged:

Standard Formula: SCORE = Right - (Wrong/(n-1)).

Right = the number of correct answers.

Wrong = the number of incorrect answers.

N = total number of alternatives per item (including the correct answer)

So, if a candidate answered 35 questions correctly on a 50-question test in which each item had 4 alternatives, the raw score after corrections would be calculated as follows: Score = 35 - (15/(4-1)) therefore Score = 30.

Alternatively, negative scoring could be employed to discourage guessing. The responses to a multiple-choice question might then be scored as follows:

+1 for the correct answer
-1 for an incorrect answer
0 for no answer.

If corrective scoring is used, it is always important to alert candidates to this, so that they can alter their test-taking strategies accordingly.


This thread stems from a question I posted on the Mensa subreddit re how guessing might inflate scores on multiple choice questions. One person said to average my scores on an IQ site and reduce by 20% Another said guessing wouldn’t make much difference to the score.

I said I guessed on tests of spatial ability.That I was clueless in working them out. Someone said

Just guessing and getting the correct answer is a “feature” of multiple choice questions. It will apply to everyone who doesn’t know the answer to every question. The person who knows 28/30 and guess on 2 will have a risk of inflating their score just as you who guess on every single question has a risk of inflating your score. If you know that you are guessing on everything then the test is meaningless for you. You are now not measuring your ability but measuring random chance. You ability is actually under what most spatial tests are capable of measuring and all you can get from them is that you are below the normal range.

I said I didn’t guess all answers but knew,or thought I knew,about 3-5 answers out of 30 ie 10-17%. Hence 83-90% were shots in the dark/random guesses.

I do think you can guess better than chance(Jukebox’s example) and worse than chance but it probably evens itself out in the end.

I tried the above mentioned formula on an IQ test I got 12 /30 questions right on with 6 choices per question.

12 - (18/(6-1)) therefore Score = 8.4 . Rounded down=8 . Feel free to correct me if my maths is wrong.

What this tells me given the high/very high amount of guessing because I’m basically clueless is that my spatial intelligence is low/very low. Indeed it is probably lower than the tests indicate because of inflation due to guessing.

If only one answer is correct among the 5 or 6 options, then you got the numbers right. However if at least one, but possibly several answers are correct, the chances of you randomly guessing right decrease dramatically.


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.