on This Site
to This Site
Uncommonly Difficult IQ Tests
Web site was created and maintained by Darryl Miyaguchi until 2004. It is now hosted here to preserve its content.
Last updated April 24, 2004
??? visitors since about Dec, 2000. About 400,000 (?) before that (starting from June 10, 1997).
Thinkers International email discussion group (Hi-IQ membership not required)
High IQ email society (Hi-IQ membership not required) [Wayback Machine link is dead]
Vitruvian Society (email list with no high IQ requirement)
High IQ Society.com 95th percentile
Mensa (USA site) 98th percentile
Sigma Society 98th percentile
Top One Percent Society 99th percentile
Mysterium 98th percentile
Intertel 99th percentile
Colloquy 99.5th percentile, email society
Poetic Genius Society 99.5th percentile
Cerebrals.com 99.7th percentile
One in A Thousand Society 99.9th percentile
International Society for Philosophical Enquiry 99.9th percentile
Triple Nine Society 99.9th percentile
Glia Society 99.9th percentile; Also, GliaWeb
IQuadrivium Society 99.9th percentile
SesquIQ 99.9th percentile
Minzar - The High IQ Society for Christians 99.9537th percentile [Wayback Machine]
Prometheus Society 99.997th percentile
Ultranet 99.997th percentile
* The Mega Society Page (West Coast) or Another site, with online issues of Noesis 99.9999th percentile
* The Mega Society Page (East Coast) and its online issues of Noesis-E. 99.9999th percentile [Vol. 1, No. 1 on Wayback Machine]
The Pi Society Page 99.9999th percentile
Mega HIQ Girls 99.9999th percentile among women [Wayback Machine]
Giga Society 99.9999999th percentile
* See this page for an explanation for multiple Mega Society links
Grady Towers was well-known and highly respected in the high IQ societies. Some of his finest contributions to the journals can be found at the Prometheus Society's website:
Theories of Multiple Intelligence
NOTE: In late November and early December of 1999, the Mega Test was compromised in high-visibility public forums, such that I no longer feel it is appropriate to host the test on this website.
The following is modified from Scot Morris's introduction to the Mega Test, in Omni magazine, Games column, April 1985, p. 128:
Most intelligence (or I.Q.) tests are designed so that average people get average scores, clustered around the midpoint I.Q. of 100. The tests are most powerful at their middle ranges, where the difference between an I.Q. of 100 and 105 may be a matter of several questions on the test itself. But at their upper ends, the tests don't seem to discriminate nearly as well -- the five-point difference between I.Q. scores of 145 and 150, say, may translate into raw-score differences of only one or two test questions.
There have been various tests devised over the years that make fine distinctions in the intellectual stratosphere. The idea is to make a test so difficult that geniuses will get average scores, and only supergeniuses will be able to achieve the highest scores.
Various high-I.Q. organizations have been established over the years. Mensa, the most famous group, is open to one person in 50 -- that is, people in the upper 2 percent of the population (98th percentile). The Top One Percent Society and Intertel have cutoffs at the 99th percentile, and the One-in-a-Thousand and the Triple Nine Societies have cutoffs at the 99.9th percentile. The Prometheus Society shoots for 1 in 30,000. But the most restrictive group is the Mega Society, which is theoretically limited to one person in a million (the 99.9999th percentile).
The founder of Mega and the author of one of its admissions tests is Ronald Hoeflin, of New York. At Omni Magazine's request, Hoeflin split his original long-form test into two parts of 48 questions each. One part, called the Mega Test, originally appeared in Omni (April, 1985, Games column). The other part appeared in the Omni I.Q. Quiz Contest, complete with answers (Marilyn Mach vos Savant, 1985; Published by McGraw-Hill Company, New York; Copyright by Omni Publication International, Ltd., and Marilyn Mach vos Savant, ISBN 0-07-039377-X).
Hoeflin estimates that the Mega Test has a floor of 100 (which means that if you get no questions right, your I.Q. is somewhere below 100) and a ceiling of 190+. At Omni's request, prior to its publication in the magazine, the test was administered to more than 150 people -- all members of the major high-I.Q. societies, in order to show, for example, that Mega members score higher than members of Prometheus, who score higher than members of Triple Nine, and so on.
1Of the test's 48 questions, 10 correct corresponds to an I.Q. of 133, the cutoff for membership in Mensa (although Mensa does not accept results of unsupervised tests like the Mega); 14 right, an I.Q. of 138, qualifies one for membership in the Top One Percent Society; 24 right, an I.Q. of 150, qualifies for the One-in-a-Thousand Society, 36 right, an I.Q. of 164, qualifies for the Prometheus Society, and 43 right, or an estimated I.Q. of 177, is the cutoff for joining the Mega Society.
One of the Omni readers who scored highest on the Mega Test was John H. Sununu, then the governor of New Hampshire, and later Chief of Staff under President Bush. His score of 44 correct gave him an estimated I.Q. of 180 (achievable by approximately one in 3 million). Marilyn vos Savant scored 46 (I.Q. of 186) on her first attempt. Two others have scored 46 on their first attempt (Note that Ron now specifies that only one attempt is allowed). About 3 people have scored 47, but only on a second attempt. This includes Eric Hart. "Eric Hart" turned out to be a pseudonym for an individual who scored 42 on his first attempt of the Mega Test [Ron Hoeflin, in a letter to me, made the distinction between first and second attempts of Eric Hart]. About 3 people have scored 45 right, including Steve Schuessler (who worked at NASA). Sununu's score of 44 has been tied by 3 or 4 other people so far, including Rick Rosner, who edited the Mega Society journal, Noesis. Of the three highest scores (46 right), two are women.
In reply to a letter I wrote to him, Chris Langan, the alter-ego of Eric Hart, has written that he sees no reason to conceal his identity any longer. It is his contention that his score of 47 could (or should) be counted as a first attempt, since his test-taking strategy was to minimize the effort and time required achieve the mega-level, which was 42 right in the original Omni magazine introduction to the Mega Test. One of my comments to him was:
"Your explanation of Eric's test-taking strategy is interesting; I suspect that few people who scored in this range followed the same strategy of maximizing the score-to-time-spent ratio with the goal of achieving a score of at least 42 correct. For those who are able, I would think there would be a powerful motivation to achieve the highest score possible, period."
His response was:
"As you speculate, the same mathematics [game or decision theory] can also be used to justify trying for the highest possible score. Mathematical expectation, as expressed by the gain/cost ratio, allows "gain" and "cost" to be defined in various ways. If one takes great pleasure in the solution of IQ test problems or in (meaningless) competition with other testees, then one might regard every minute spent on such a test as emotionally "gainful." And if one has nothing better to do anyway, the cost to one's other pursuits may be small. This can result in a higher perceived gain/cost ratio for solving as many problems as possible, at least for idle puzzle addicts."
1 The raw score equivalent IQ's are from Ron Hoeflin's sixth norming. The original column by Scot Morris read as follows:
Of the test's 48 questions: 8 correct corresponds to an I.Q. of 134, the cutoff for membership in Mensa; 22 right, and I.Q. of 150, qualifies one for membership in the Triple Nine Society; 33 or above, corresponding to a 164 I.Q., qualifies one for membership in the Prometheus Society; 42 right, or an estimated I.Q. of 176, is the cutoff for joining the Mega Society.
The Titan Test is a more difficult twin to the Mega Test. It consists of the best problems from six preliminary tests that were tried out on volunteers between 1985 and 1988. The May 14, 1997 issue of The Wall Street Journal 2 reports that "Rick Rosner, a TV writer, bar bouncer, nude model [is] the only person to get a perfect score on Dr. Hoeflin's latest test [I am informed that the test referred to is the Titan, which is no longer the latest test -- DTM]."
2 Visit the "More Reading Material Page" for summary information about this article.
Third in the series of tests that Ron Hoeflin has designed. First indications are that this test has a floor of about 100 I.Q. and a ceiling of about 180, which makes it slightly easier than the Mega or Titan and accessible to a wider group of people.
Combines the best problems from the Mega, Titan, and Ultra Tests, while omitting the verbal analogies and number sequences. Non-native English speakers may thus prefer this test to the others (and it can be more readily translated, if needed). Also, this test provides a legitimate avenue for resubmitting answers to problems on the other tests (as long as the test taker believes the Hoeflin Power Test is still a valid measure of his intelligence). So far (as of January 2001), the high scorer on the Power Test has been David Michael Fabian of Buffalo/Houston.
04/24/04 Chris Cole wrote to me: "As you know the Mega Society has been looking around for a new entrance test now that the Mega Test has been compromised. Were currently using the Titan Test but I fear that it will also be compromised in the future. Accordingly, some members of the Society have been working on a cheat-resistant test that can be administered online. It has several cheat-resistant features, one of which is that it never asks the same question twice. It is currently in a very preliminary stage, but you can take a look at it at www.mental-testing.com. The test is adaptive in the sense that it adjusts your score as more people take it. This is because the estimated difficulty of the questions is changed as more people take the test. I wonder if youd be willing to help us get some sample test takers by featuring it on your site? 04/12/03 Ron writes: "I am no longer going to score the Ultra Test or the Hoeflin Power Test, but am still willing to score the Titan Test and Mega Test. I am lowering the scoring fee for the latter two from $30 to $25, payable to 'Ronald K. Hoeflin.' Those requesting a copy of the Mega Test should send $25 with their request for the copy rather than waiting to pay if and when they send their answers. The Mega Society, One-in-a-Thousand Society, and Top One Percent Society all accept the Titan Test for admission purposes but only the latter two societies accept the Mega Test for admission." 09/24/01 Ron writes: "The lower scoring fee of $20 for my tests does not seem to be stimulating much interest, so please raise the fee to $30 per test." 07/30/01 My email address is changing to "miyaguch at eskimo dot com" 05/20/01 Tommy Smith has retired the OCIP 04/07/01 Added Mega Test Score Interpretation Booklet (first norm) 04/05/01 Ron Hoeflin asked me to lower the advertised fee on each of his tests to $20. 03/05/01 I am reducing my activity on this site to bare minimum. 02/25/01 Added link to Brain Board 3 02/23/01 Added link to BrainBoard Continued's new location 02/23/01 Changed link to Glia brochure; added link to GliaWeb 02/23/01 Paul Cooijmans has a new email address 02/02/01 Added link to Bill Bultas's Difficult Analogies test 02/02/01 Added link to Noesis-E 02/02/01 Added link to Mega HIQ Girls 01/24/01 Luminary Society appears to be gone (Wayback Machine link) 01/18/01 Added Laurent Dubois's 9 I 6 01/18/01 Paul Cooijmans has an email address now 01/08/01 Added link to Vitruvian Society 01/08/01 Added link to Inner Sanctum 10/31/00 Added link to an additional Mega Society Page (with online issues of Noesis) 10/29/00 Added link to Nathan Haselbauer's NYC High IQ Society Membership Exam 10/28/00 Added link to Nicolas-Elena's Tests 10/28/00 Added Robert Lato's Logima Strictica 36 10/14/00 Paul Cooijmans is feeling better 10/13/00 Moved "Beta Test" information to the Tests section 10/13/00 Added link to UltraHIQ 10/07/00 Added Beta test Rasch analysis 10/06/00 Added Beta test score comparability with SAT 10/06/00 Added Greg Grove's Mysterium 10/06/00 Added 1966-68 ACT information 09/26/00 Added Beta test factor analysis 09/12/00 Added Beta test discrimination plots 09/10/00 Added Beta test correlation with SAT 08/22/00 Added Norming #2 of the Langdon Adult Intelligence Test 08/22/00 Added SesquIQ site 08/22/00 Added the Marilyn is Right! page 08/05/00 Added Easy Factor 2.5 and Rasch 1.0 programs 08/02/00 Added Bob Seitz' Vocabulary Page 07/31/00 Added Bob Seitz' page 07/30/00 Added Supplementary Statistical Data for the LAIT, Norming #2 07/29/00 Copied Herb Weiner's "Marilyn is Wrong!" site link to the main page. 07/29/00 Added link to Thomas Lisa's song, "Questions for Marilyn" 07/29/00 Added results of Power Test (46 scores) 07/29/00 Added Grady Towers's "The Empty Promise" 07/29/00 Added posthumous letter from Grady Towers -- my last from him 07/29/00 Added link to Scientific American's special issue on intelligence 07/23/00 Removed tests by Paul Cooijmans at his request. Paul is experiencing health problems and wishes to cut down on his incoming mail. 07/20/00 Added link to Queendom.com's "Culture-Fair IQ Test" 07/20/00 Added link to Nik Lygeros' M-classification 07/06/00 Removed Greg Grove's Ready Reckoner at his request 06/27/00 Added direct link to Xavier Jouve's "Encephalist IQ T1 revised" test. 06/27/00 Added link to the Luminary Society 06/04/00 The Giga Society now has an official web site 06/03/00 Added link to Geniusweb: (Purported) Geniuses on the Web 06/03/00 Added link to High IQ email society 06/03/00 Added link to the Minzar Society (for Christians) 05/30/00 Re-added link to Tommy Smith's and Joel Willis' OCIP, which is hosted on the Cerebrals.com site. 05/30/00 Added link to tests by Xavier Jouve 05/23/00 Added link to "The IQs of Eminent Geniuses" page, with corrections for the Flynn Effect. 05/23/00 The Poetic Genius Society continues with approval from Greg Grove 04/30/00 Added Paul Cooijmans's Test for ESP (with statistics) 04/26/00 Linked to Grady Towers's "Theories of Multiple Intelligence" 04/25/00 Greg Grove has discontinued the Poetic Genius Society 04/18/00 Tommy Smith has closed MentalMadness. 04/09/00 Added link to the Sigma Society. Thanks to Barry Howard for the information. Barry's High IQ site [Wayback Machine] 03/25/00 Added followup to Grady Towers's "On using multiple tests for high IQ society admissions." 03/01/00 Added link to Justin Chapman's page (who is 6.5 years old). 03/01/00 Added 1999 ACT national and state scores. Thanks to Mike Hess for the link. 03/01/00 Added data from 1972-73 GRE. Thanks to Mike Hess 02/20/00 Added The Final Test Statistical report, June 1999 02/19/00 Added TFG Short Form Statistical report, June 1999 02/16/00 Added link to SAT/ACT comparison (College Board). 02/14/00 Added Mental Madness's Online Store 02/12/00 Removed Boffin Board link (http://dewell.com/boffinboard/) (has been broken for a long time). 02/01/00 Linked to Paul Maxim's website 01/25/00 Added link to Kevin Langdon's Polymath Systems Home Page (still in skeleton form). 01/22/00 Added Enigmus's Discussion Forum; Perception Test removed and replaced with The Chaos 01/19/00 Added link to Cerebrals.com 01/19/00 Changed link to Mega Society Page (West Coast). Thanks once more to Barry Howard for the information. Barry's High IQ site [Wayback Machine] 01/18/00 Added results of Hoeflin's "Smartest Person in the World Contest" 01/11/00 Added explanations for the compromise and removal of the Mega Test. 01/11/00 Paul Cooijmans revised the TFG (long form). 01/06/00 Added Peter Schmies' Quantity Quiz 01/06/00 Added link to the Mega Foundation's Ultranet (Global Ultra HiQ Network). 01/03/00 Paul Cooijmans dropped his idea for recruiting new testers.
View Change History prior to Jan. 1, 2000