The ACE Score combines three sub-scores: Authority, how total and coercive the leader’s power was; Cult, how devoted to and dependent on the leader followers became; and Exit, how hard it was to leave. Higher scores indicate a tighter grip.
The score is an index, not a percentage or a probability. An ACE score of 72 puts a system high on the scale, but it does not mean that anything is 72 per cent true or 72 per cent likely. We built it on the framework. The method below explains its parts, and the formulas are in the technical notes.
AuthorityTruth, Right to live
+
CultDebt, Dependence, Fear of the end
+
ExitFamily, Freedom to leave
÷ 3 =
ACE Score0 to 100
ACE is a study tool examining authority, commitment and the cost of leaving. Inclusion in ACE does not itself classify a leader or movement as a cult. Closure–Debt–Vector is this project’s synthesis of scholarly ideas, rather than an established or independently validated academic scale.
Methodology
01
Scored 0 to 10
How sealed is the bubble around followers?
Metaphorically speaking, leaders can “build walls” around the people who follow them. We look at five of these walls, and the Bubble Score is their average height. The first is control over information. When the leader is the only source of news, followers have no way to check what they are told. The second cuts followers off from family and friends outside the group, so nobody is left to pull them home. The third makes it physically hard to leave, through a remote compound, confiscated passports or constant watching. The fourth is fear of the end. If an apocalypse or a final victory always seems close, a life outside the group stops making sense. The fifth wall is the group’s claim to decide who deserves to live, and once a group crosses that line, leaving can become dangerous.
Think of the difference between a submarine and a town square. In a submarine every hatch is shut. In a town square, anyone who feels uneasy can simply walk off.
02
Scored 0 to 10
How much do followers feel they owe the leader?
Over the years, followers build up a debt to the leader. Some of it is material, such as money, houses and savings handed over. Some of it is emotional: confessions the leader keeps on file, careers and relationships given up, or the belief that the leader saved their life. The bigger the debt, the more leaving feels like admitting that all of it was wasted. In many groups this can hold people as firmly as a locked door.
It may also help explain why disappointment does not always end loyalty. Leon Festinger and colleagues proposed that people who have invested heavily in a belief, and who have the support of fellow believers, may reinterpret a failed prophecy rather than admit the investment was a mistake. Later studies found that many believers instead lose faith or leave, so this depends on circumstances.
03
Scored 0 to 10
How far does everything rest on one person?
Max Weber pointed out that power built on one person’s charisma is fragile. A movement that wants to outlive its founder has to turn that charisma into offices, rules and a way of choosing a successor. Leader Dependence measures how far that has failed to happen. A high score means the system has few institutions of its own and would fall apart without the leader. A low score means it has real structures that could carry on without them.
Institutions also matter to ordinary followers. A rule can be appealed and an official can refuse an order, and each of those openings makes it a little easier to get out.
04
Scored −10 to +10
When the system turns violent, where does the harm go?
When a closed system turns violent, the violence goes somewhere, and this score records which way it went. It records direction, not scale: the size of the harm is described in each profile’s evidence and caveats. A negative score means it turned inward, against the group’s own members. A positive score means it turned outward, against enemies, minorities or anyone labelled an outsider. A score near zero means there was little lethal violence, or about as much in each direction.
Jonestown is the clearest inward case, since the people who died were the members themselves. Nazi Germany is the clearest outward case. The regime murdered millions of people it had defined as outsiders.
Reading a score
04710
0 to 3, low
Control is light. Followers keep ordinary lives outside the group and can leave without losing much.
4 to 6, medium
Followers are under real pressure, but family, work and outside institutions still offer a way back.
7 to 10, high
Leaving has become expensive, frightening or physically impossible.
Ending patterns
: four rule-based labels
Each profile gets one of four pattern labels from a fixed rule (see the technical notes). The labels describe a combination of scores. They are not forecasts, and they have not been tested against independently coded historical outcomes, so a case can carry a label that differs from how it actually ended. Each profile’s summary and caveats describe the documented outcome.
A closed system with little institutional restraint, where harm was aimed mainly at its own people.
Jonestown and Heaven’s Gate fit this pattern. Waco carries the same label, although it ended in a government siege whose causes are disputed.
A closed system with little institutional restraint, where harm was aimed mainly at outsiders.
Nazi Germany fits this pattern and was defeated from outside in 1945.
A closed system with strong offices, rules and a way to choose a successor.
The Soviet state after Stalin is a common illustration, although Stalin’s own scores put him just under the threshold.
Walls low enough (a Bubble Score under 5) for followers to leave fairly easily.
A low score does not rule out a violent end: Canudos carries this label and was destroyed by the Brazilian army in 1897.
Scoring guide: what each score means
Every profile, in the database and from the live Model Engine, is scored against these anchors. Scores between the anchors are allowed. Jim Jones and Adolf Hitler are used as fixed reference points so that scores stay consistent across cases.
Who decides truth (0 to 10)
0Free access to outside information; internal dissent tolerated.
3The leader’s interpretation is privileged; outside sources discouraged but available.
5Outside information systematically denigrated; internal teaching dominates; doubt treated as a personal failing.
7Outside media, mail or speech censored; doctrine treated as infallible; dissent punished.
10Total information control backed by terror, or a sealed setting with constant indoctrination.
Cut off from family (0 to 10)
0Ordinary family and friendship ties untouched.
3The group takes priority, but outside ties are tolerated.
5Outside relationships discouraged; marriage, child-rearing or friendships steered by the group.
7Relatives and outsiders labelled enemies; contact restricted; family structures reorganised.
10Near-total severance: children separated from parents, members inform on family, outside contact forbidden.
Freedom to leave (higher means harder to leave) (0 to 10)
0People come and go freely.
3Communal living common but voluntary; leaving is easy.
5Group housing or compounds; practical barriers such as money or documents held by the group.
10Armed guards or shoot-to-kill borders, confiscated passports, remote enclave with no practical exit.
Fear of the end (0 to 10)
0No apocalyptic, millennial or destiny framing.
3Vague utopian goals on a long horizon.
5Strong millennial expectation or historical-destiny ideology that shapes daily life.
7Imminent catastrophe or final struggle repeatedly predicted; life choices subordinated to it.
10A dated or immediate end, rehearsed in drills or preparations.
Right to live (0 to 10)
0No dehumanisation of outsiders or dissenters.
3Outsiders described as lost, impure or inferior.
5Enemies dehumanised in doctrine; critics harassed or threatened.
7Violence against defectors or enemies sanctioned, with deaths documented.
10Systematic killing or mass death.
Devotion Debt (0 to 10)
0Few demands on members.
3Regular tithes, time commitments, moderate sacrifice.
5Large financial and personal sacrifice; confession or self-criticism practices.
7Property, careers or family roles surrendered; confessions used as leverage; livelihood depends on the group.
10Total dependence: all assets surrendered, blackmail material held, survival depends on the leader.
Routinisation (shown on the page as Leader Dependence = 10 minus this score) (0 to 10)
0Everything depends on the leader personally; no succession plan.
3Some offices exist but have no authority apart from the leader.
5A functioning organisation with officials; succession unclear.
7Strong party, state or church bureaucracy; a recognised succession mechanism.
10Fully institutionalised; the leader is replaceable by procedure.
Who gets hurt (−10 to +10)
−10Mass death mainly of members.
−5Serious violence against members, little against outsiders.
0Little lethal violence, or roughly balanced.
+5Serious violence against outsiders, such as terror attacks or persecution.
+10Genocide or wars of aggression against outsiders.
Technical definitions and formulas
We built the ACE Score on Closure–Debt–Vector (CDV), our own synthesis of Lifton’s eight criteria of thought reform, Weber’s types of authority and Goffman’s total institution. Lifton describes the walls and Weber explains where the leader’s power comes from. CDV adds the direction of harm and an estimate of what it costs to leave. The list below matches each plain name on the page to the scholarly idea it adapts. The five walls, the numerical anchors, the weights and the thresholds are this project’s own operational choices; the scholars named here did not propose these measures or score these cases.
Totalism Index: our own measure, inspired by Lifton’s account of totalism, computed as the mean of the five closures. Who decides truth is epistemic closure, which combines Lifton's milieu control and sacred science. Cut off from family is social closure, the severing of kinship and outside ties. Freedom to leave is spatial closure, covering geography, movement and the body in Goffman's sense, and it is scored so that a higher number means less freedom. Fear of the end is temporal closure, the future foreclosed by apocalypse or destiny as described in Cohn's work on millenarianism. Right to live is existential closure, Lifton's dispensing of existence.
Commitment Lock: our own measure, adapting an idea from Kanter's Commitment and Community (1972), which treats accumulated obligation in the form of confessions held as leverage, sunk sacrifices, sacralised gifts and dependency for survival as the mechanism that converts dissonance into loyalty rather than exit.
Personalism, displayed as ten minus Weber's routinisation score. The model still reports routinisation, that is the conversion of charisma into traditional or legal-rational structures, and the page flips it for display so that a higher number consistently means a more fragile and more leader-centred system.
Violence Orientation, formerly the annihilation vector, running from −10 for violence turned entirely inward to +10 for violence turned entirely outward.
The average of three sub-scores, each from 0 to 100. Authority is (epistemic + existential) ÷ 2 × 10, our formula, drawing on ideas from Weber and Arendt. Cult is (debt + (10 − routinisation) + temporal) ÷ 3 × 10, drawing on Weber, Kanter and Cohn. Exit is (social + spatial) ÷ 2 × 10, drawing on Hirschman’s idea of the cost of exit. The index is not a percentage or a probability. Version 1.0 used Closure × Debt ÷ (1 + Routinisation ÷ 10), which measured lock-in only; version 1.1 measures the three parts named in ACE directly. Averaging three sub-scores is itself a weighting choice: each Authority or Exit component counts for one sixth of the ACE Score, while each Cult component counts for one ninth.
Ending pattern, formerly called Collapse Mode, assigned by rule and descriptive only. Closure below 5 gives Dissipation; otherwise routinisation of 6 or more gives Ossification; otherwise a vector of +2 or more gives Confrontation, and anything remaining is Implosion. The two anchors on the map are Hitler at closure 8.6 and vector +10, and Jones at closure 9.4 and vector −8.
ACE is a study tool examining authority, commitment and the cost of leaving. Inclusion in ACE does not itself classify a leader or movement as a cult. Closure–Debt–Vector is this project’s synthesis of scholarly ideas, rather than an established or independently validated academic scale.
Limitations and corrections
ACE is an exploratory, interpretive framework. Read its scores with these limits in mind.
Retrospective scoring of a selected set. The 100 cases were chosen and scored after the fact. Patterns across them describe this dataset, not leaders or movements in general.
Some patterns are partly built in. Right to live, one of the five walls in the Bubble Score, rises with realised killing, so the link between closure and violence on the closure map is partly circular. It is not independent evidence that closure causes violence.
Direction is not scale. Who gets hurt records where violence was aimed, not how much there was. A score near zero can mean little violence or heavy violence in both directions.
Who counts as an insider varies. Profiles sometimes count people a regime excluded from its national community as outsiders even when they were citizens. Each profile’s caveats say which deaths fall on which side.
Ending patterns are labels, not forecasts. They come from a fixed rule that has not been tested against independently coded outcomes.
Evidence is cited at profile level. Each profile lists its key sources, but individual scores do not yet carry page references, and no inter-rater comparison has been published. Small differences in rank, such as 81 against 83, should not be over-read.
Corrections log
30 October 2026. Pol Pot: corrected “died without trial” to record his 1979 conviction in absentia and the 1997 Khmer Rouge proceeding. Adolf Hitler: separated the 1936 Hitler Youth law from the 1939 compulsory-membership decree. Method: the Hitler anchor now reads from the profile data (vector +10, not +8.5). Ending patterns are now presented as rule-based labels rather than predictions. Festinger’s findings are described as conditional.