AVRELIVS
Join the waitlist
AVRELIVS
Join the waitlist

The method

Evidence before intelligence — the method behind Avrelius.

Change careers or start your own company, move or stay, choose the person to build a life with — decisions like these shape decades. Yet we sometimes base them only on how we now remember and interpret past experience, other people’s opinions, and the mood of the moment. At the decisive moment, there is often no other evidence to rely on.

We created Avrelius to give each of us firmer ground. It turns lived experience into a coherent record, connecting notes, decisions, expectations, and actual outcomes over time. AI helps reveal patterns and blind spots, surface possible tensions, add knowledge from other fields, and suggest the next step.

This article explains the method, the research behind its five principles, and where the method applies. Every scientific claim can be checked against the references at the end.

§ 01 — Familiar approaches

Why is one approach not enough?

Memory, journaling, willpower, and outside advice can help us decide more deliberately. But none alone provides a complete and reliable basis for a decision. Below, we cite research demonstrating the limitations of these familiar ways of making decisions and examine what they have in common.

Five paths leave the same spot: memory, journaling, willpower and advice each end at a marker naming why they stop; the evidence path runs on over the hill.

Relying on memory and self-analysis

Once we know how a situation turned out, the mind quietly reshapes our memory of what we expected. The outcome begins to seem obvious in hindsight, though we may have judged the situation differently before learning it. This well-documented effect is called hindsight bias1,2,38.

Introspection also has limits. When explaining our actions, we often construct a plausible account after the event because we lack direct access to every internal process that shaped the decision3. In comparative judgments, people often rate themselves above others4. Self- and outside assessments of personality often agree, but their accuracy varies by trait and context5,41,49. Memories and explanations formed after the fact therefore do not provide a complete or impartial picture.

Keeping a journal without analyzing it

Notes accumulate, but without systematic analysis they rarely form a coherent picture6. Another risk arises when someone repeatedly returns to the same failures without reinterpreting them: a journal can sustain rumination — repetitive negative thinking that increases distress7.

In a controlled study, participants who wrote only about their emotions after a stressful event felt worse than those who also reflected on its causes, meaning, and consequences8.

Without that reflection, a journal can help us see the maze of our own thoughts, but not always the way out.

Counting on intention and willpower alone

A firm intention shows what someone wants but weakly predicts behavior change. Even when researchers substantially strengthened participants’ intention to act, actual behavior changed far less9.

A goal begins to guide action when the desired outcome is clear and progress is checked regularly10. Promises to ourselves — “I’ll start,” “I’ll stop,” “I’ll change” — are therefore not enough: we need concrete actions and a way to check their results.

Misjudging advice

We often give our initial opinion too much weight and underrate outside advice — even when the other person is more experienced and better informed11. With AI, the risk reverses: a persuasive answer is easy to mistake for an independent, unbiased assessment39.

Language models are trained partly on human preferences. Because users more often reward answers that confirm their position, models receive a systematic signal to agree12,40. In a study of 11 modern models, they affirmed users were right about 50% more often than people did13. In follow-up experiments about real personal conflicts, participants who spoke with an agreeable model became more convinced they were right and less inclined to repair the relationship — yet rated its answers more highly, trusted them more, and were more willing to return13.

Advice that always confirms we are right comes at a cost: the decisions we make based on it.

Conclusion

These approaches share one weak spot: the quality of the feedback we usually get from our own experience.

For experience to teach us, we need the full chain: decision — expected result — confidence in the prediction — actual outcome. In ordinary life, it almost always remains incomplete: consequences arrive months or years later, outcomes can have several causes, initial expectations are rarely recorded, and memory reshapes them to fit the known result.

Decision science calls this a “wicked” learning environment: feedback is delayed and ambiguous, with no objective benchmark for the result. We can therefore face similar situations repeatedly and still learn the wrong lessons from experience14,15.

Avrelius does not replace personal experience: that experience remains the system’s main raw material. Avrelius changes how we learn from it by closing this chain and making it transparent. The method’s five principles show how.

Two decision cycles: in ordinary life, the link between a decision and its consequences breaks; in the Avrelius cycle, the decision connects to a prediction, stated confidence, a review date, the actual outcome, and the conclusion drawn afterward.

§ 02 — Five principles

How to turn personal experience into evidence-based feedback.

Avrelius combines five principles into one system, each offsetting the others’ limitations. It preserves and compares your notes about events, thoughts, feelings, intentions, actions, and decisions, looking for recurring patterns and possible tensions. Every conclusion rests on specific facts you can return to and check.

A fact here is not a judgment about a person but a recorded part of their history: what was said, when, and under what circumstances. “I work well under pressure” does not prove someone is effective under stress; only that they described themselves that way on a particular day. A broader conclusion requires comparison with other notes, actions, decisions, outcomes and, with the user’s consent, other people’s observations. Avrelius therefore starts with facts, not plausible guesses or assumptions.

Evidence before interpretation

Avrelius draws a conclusion only when it can show the data behind it. For now, your notes are the main source: the system preserves the original wording, date, and context. If you said, “After the team meeting, I felt that I was avoiding conflict again,” Avrelius cites that entry rather than turning it into the impersonal claim “You avoid conflict.”

In the future, with your permission, the model may also use correspondence, wearables, questionnaires, and input from people you invite. Because these sources differ in nature and reliability, they cannot be interpreted alike.

When information is insufficient or sources conflict, the system says what it does not yet know and offers a testable hypothesis rather than presenting a guess as knowledge. This reduces the risk of overconfidence16.

Triangulation over introspection

We know our own intentions, doubts, and explanations. Others see our recurring behavior, the consequences of our decisions, and how we act in relationships and at work. Neither view is complete: self-analysis is more accurate for some aspects of personality, outside assessments for others, and combining sources is usually more reliable than either alone5,17,41.

Avrelius therefore compares your words with your decisions, outcomes, and changes over time. Later, if you wish, the picture can include feedback from people you choose.

Repeated gaps between words and actions are especially important. Someone may, for example, write several times that they want to work less — and keep taking on new projects. A single discrepancy is not enough: behavior naturally changes with circumstances18. Repeated patterns can signal possible tensions. Avrelius raises them as questions and cites the entries behind them.

Three data streams — the user’s notes, decisions and outcomes, and changes over time — form a portrait; areas without enough information remain unfinished.

Score the judgment, not the person

Avrelius evaluates not you as a person, but the accuracy of your predictions and the quality of your decision process. Before an important decision, you record the expected outcome, your confidence, and a review date: “I estimate a 70% chance that the new project will break even within six months.” Six months later, you record the actual result.

One outcome says nothing about a prediction’s accuracy: roughly three out of ten predictions made at 70% should fail. Calibration can be assessed only across predictions with similar stated probabilities. Once there is enough data, Avrelius shows how closely confidence matches actual outcomes across areas of life. Until then, no statistics are shown: small-sample percentages depend too much on chance19,20.

Regularly recording probabilistic forecasts, checking them, and reviewing errors improved judgment accuracy in the largest experiments on geopolitical forecasting21,22,42. Showing outcomes or scores alone does not reliably improve calibration50. Avrelius therefore applies the full cycle to personal decisions: recording the forecast, checking the outcome, and actively reviewing the error.

A well-founded decision can still fail through unforeseen circumstances, while a weak one can succeed through luck. Judging a single decision by its outcome alone therefore falls prey to a documented bias23. Avrelius instead considers the initial evidence, reasoning, and actual outcome separately.

The chart compares stated confidence with how often predictions came true: if ten predictions were rated at 90% and six came true, the actual frequency is 60%.

Distance, then direct

Personal involvement keeps us from reasoning about our own problems as even-handedly as we do about someone else’s. Experimental studies suggest that self-distancing — mentally viewing ourselves from the outside — can improve reasoning about personal conflicts and support emotional regulation in some situations24,25,43. Evidence from daily life indicates that the benefit depends on the situation51.

Avrelius uses this approach in the weekly review: first it shows the pattern and dated entries, then explains what the pattern may mean and proposes a concrete next step. This places your words, decisions, and actions in the context of the current situation.

Excerpt from a weekly review

Observation

Since April, you have called your current job “temporary” four times. This week, you agreed to lead a project planned to run for at least a year.

Reflection

Your words point toward leaving, while this new commitment assumes you will stay. Delaying the choice may feel safer — but over time, each new commitment narrows it.

Action

Name the date by which you will decide to stay or to go — and put it in the calendar this week.

This format rests on feedback research. Feedback improves performance on average but makes it worse in about a third of cases, especially when it shifts attention from the task to self-evaluation26,44. More effective feedback answers three questions: what the person is aiming for, where things stand, and what to do next27. Avrelius therefore examines concrete actions, not the person.

The next step must be concrete45. A plan such as “when X happens, I will do Y” roughly doubles follow-through compared with a general intention28,46. Research on deliberate personality change likewise finds that change comes through small actions aligned with the chosen goal29,47.

A conclusion’s directness depends on the volume and quality of the data. Early on, Avrelius asks more questions and offers more hypotheses. A confident conclusion is warranted only when supported by several entries from different times or by different sources. The same rule applies to criticism and praise.

Autonomy over compliance

Avrelius does not use guilt, shame, or fear of losing progress to bring you back to the app. After a pause, you continue where you left off.

Motivation based mainly on outside pressure, guilt, and shame is less durable and more often leads people to abandon what they started30,48. When missed days are made visible, a broken streak can reduce the desire to continue more than having no streak display at all31. After a setback, treating ourselves with kindness helps us return to a goal better than self-criticism does32.

Autonomy also applies to the system’s observations. You can correct the information behind the digital portrait, confirm or reject a hypothesis, add missing context, or choose not to follow a recommendation. The decision remains yours; Avrelius shows the basis for a conclusion rather than demanding compliance.

The five principles complement one another: structured records prevent repetitive thinking; the next step turns insight into action; an outside view reduces self-reproach; comparing entries and sources keeps a chance discrepancy from being mistaken for a stable pattern; and requiring enough evidence keeps a hypothesis from being presented as knowledge.

Their combination is the method.

Five elements form a closed ring: a source for every conclusion, multiple-source comparison, predictions checked against outcomes, an outside view followed by a concrete action, and user autonomy.

§ 03 — Where it comes from

An ancient practice, with modern instruments.

Marcus Aurelius wrote the Meditations for himself, systematically examining his own actions, judgments, and reactions33. Avrelius continues this Stoic tradition of systematic self-observation.

Some Stoic exercises have modern analogues studied in experimental psychology: evening review resembles structured self-monitoring; the “view from above,” self-distancing24; and mental preparation for obstacles, mental contrasting — imagining a desired future alongside the specific obstacles in the way34.

The connection between Stoicism and modern psychology is well documented. Albert Ellis and Aaron Beck, the founders of cognitive therapy, explicitly named the Stoics as a philosophical source of their work35,36. Their approaches grew into one of the most extensively studied families of psychological methods37. This lineage explains the origins of some Avrelius practices, but it does not make the product therapy.

The modern part of the method also draws on decision science: once enough predictions have been compared with outcomes, Avrelius shows how closely stated confidence matches the actual frequency of outcomes21,42.

A timeline of the method: from Marcus Aurelius’s Meditations around AD 170, through the Stoic roots of cognitive therapy and twentieth- and twenty-first-century decision research, to the Avrelius method in 2026.

§ 04 — What’s different

How the Avrelius method differs from an ordinary AI advisor.

We won’t name specific products. The differences are in the approach itself.

A structured model, not isolated memories

An ordinary AI assistant’s memory stores isolated details from past conversations. Avrelius builds a structured model the user can inspect. For every observation, it stores the source, date, related areas of life, current relevance, and supporting and conflicting evidence.

The model changes with the person: new entries refine an earlier hypothesis or show that it is outdated. When data is insufficient, Avrelius leaves the gap open rather than filling it with plausible text. A system that ignores the limits of its knowledge is not modeling a person — it is improvising, or simply lying.

Grounded conclusions, not agreement to please

The tendency of language models to agree, discussed above, is built into their training12,13,39,40; a single instruction to “be honest” cannot correct it.

Avrelius counters this risk through the response mechanism, not a tone instruction alone. Every conclusion must cite specific entries and match how much is known about the user. Criticism must address observable behavior, not personality, and end with a proposed action. Praise follows the same standard: unsupported approval is as much an error as unsupported criticism. Yet personal context can itself increase sycophancy39. The mechanism must therefore preserve evidence-based disagreement as the record grows. Honesty is a requirement to test, not an assumed property.

Your experience, supplemented by AI knowledge — not generic advice

Generic advice ignores a particular person’s history. Avrelius first analyzes your notes, decisions, expectations, and outcomes, then adds knowledge from psychology, business, finance, and other fields to clarify the situation, compare options, and identify risks.

General knowledge and AI assumptions are not presented as facts about you. A recommendation shows what comes from your experience and what from outside knowledge and model interpretation, letting you assess each conclusion’s basis yourself.

Generic advice is easy to dismiss. Your own words, with dates and context, are harder to reject — and their source can always be checked11.

A closed loop, not an open conversation

A conversation can produce any number of opinions. But judgment can be tested only through a closed loop: prediction — actual outcome — evaluation. Over time, this creates a record of decision quality that no amount of eloquence can replace14,21.

Beside a fragment of memory lies a connected record in which every claim has a source, date, area of life, current relevance, and supporting evidence.

§ 05 — Honest limits

What we do not claim.

We apply the same rigor to user data and our own promises, distinguishing what research supports from what remains a product hypothesis.

Avrelius is not scientifically proven to improve your life or decisions

Avrelius has not yet been scientifically shown to improve users’ lives or decision quality.

Individual mechanisms in the method — structured reflection, self-distancing, feedback, planning, and training in more accurate probability judgment — have been studied scientifically. Their combination within Avrelius has not yet undergone randomized controlled trials.

We can say the product’s design rests on published evidence, but cannot transfer effect sizes from individual studies to Avrelius or treat its benefit as proven.

The documented effects of the individual practices are mostly modest

In the scientific literature, the effects of structured reflection, feedback, and planning are usually small or moderate and emerge after several weeks of practice. Study averages do not mean everyone will experience the same effect.

We hypothesize that combining several evidence-based practices with regular app use will produce a cumulative effect. Until the product itself is studied, this remains a hypothesis; we therefore promise no rapid, radical, or uniform changes.

Avrelius is not therapy

Avrelius does not diagnose, treat, or replace a psychologist, psychotherapist, or physician. It is intended for structured self-reflection and decision support. Nor does it provide legal advice, regulated investment advice, or emergency services. If you are struggling to manage everyday life, experiencing acute emotional distress, or concerned about your mental health, seek help from a qualified professional.

We do not claim the AI “knows you better than you know yourself”

Avrelius can weigh more entries at once than a person can hold in mind when deciding. But more data does not make the system all-knowing or eliminate interpretation errors.

Avrelius helps you notice connections and test your assumptions — not pass judgment on you.

Citing a study does not mean its authors are affiliated with Avrelius

The studies in the References section describe the mechanisms, limitations, and historical connections used to develop the method.

The cited researchers did not help create Avrelius and are not affiliated with it. Citing their work does not imply endorsement or recommendation.

How we will test our own promises

During beta testing, we plan to test three parts of Avrelius separately: whether active error review improves calibration beyond showing outcomes or scores; whether conclusions about the user match predefined, domain-specific evidence rather than only the user’s agreement; and whether the same model preserves evidence-based disagreement with no history, a short history, and the full personal context. We will track confirmations and corrections separately: they show how users assess the portrait, but do not by themselves prove that it is accurate.

We will use pre-designed scenarios and data from participants who separately consent to the research. Alongside the results, we will publish the methodology, sample size, and limitations. Until enough data exists, Avrelius’s effectiveness cannot be considered proven.

§ 06 — Your data

Your data, in one paragraph.

Avrelius works with personal records, so users retain control of their data. It is used only to run the product and build the digital twin, never for advertising, and third-party AI services may not train their models on it. Each task receives only the text or audio fragment it needs. Original audio is stored for seven days after successful processing or an error; transcripts and derived records remain in a downloadable archive. A deletion request disables the account immediately; seven days later, all related data and files are permanently deleted. Full terms and the current list of external providers are on the privacy page.

How your data is protected →

§ 07 — References

The research behind the method.

  1. Fischhoff, B. (1975). Hindsight ≠ foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1, 288–299.
  2. Fischhoff, B., & Beyth, R. (1975). “I knew it would happen”: Remembered probabilities of once-future things. Organizational Behavior and Human Performance, 13, 1–16.
  3. Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231–259.
  4. Zell, E., Strickhouser, J. E., Sedikides, C., & Alicke, M. D. (2020). The better-than-average effect in comparative self-evaluation: A comprehensive review and meta-analysis. Psychological Bulletin, 146, 118–149.
  5. Vazire, S. (2010). Who knows what about a person? The Self–Other Knowledge Asymmetry (SOKA) model. Journal of Personality and Social Psychology, 98, 281–300.
  6. Kersten-van Dijk, E. T., Westerink, J. H. D. M., Beute, F., & IJsselsteijn, W. A. (2017). Personal informatics, self-insight, and behavior change: A critical review of current literature. Human–Computer Interaction, 32, 268–296.
  7. Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3, 400–424.
  8. Ullrich, P. M., & Lutgendorf, S. K. (2002). Journaling about stressful events: Effects of cognitive processing and emotional expression. Annals of Behavioral Medicine, 24, 244–250.
  9. Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132, 249–268.
  10. Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57, 705–717.
  11. Bonaccio, S., & Dalal, R. S. (2006). Advice taking and decision-making: An integrative literature review, and implications for the organizational sciences. Organizational Behavior and Human Decision Processes, 101, 127–151.
  12. Sharma, M., et al. (2024). Towards understanding sycophancy in language models. International Conference on Learning Representations.
  13. Cheng, M., Lee, C., Khadpe, P., Yu, S., Han, D., & Jurafsky, D. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science, 391, eaec8352.
  14. Hogarth, R. M., Lejarraga, T., & Soyer, E. (2015). The two settings of kind and wicked learning environments. Current Directions in Psychological Science, 24, 379–385.
  15. Kahneman, D., & Klein, G. (2009). Conditions for intuitive expertise: A failure to disagree. American Psychologist, 64, 515–526.
  16. Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological Review, 115, 502–517.
  17. Connelly, B. S., & Ones, D. S. (2010). An other perspective on personality: Meta-analytic integration of observers’ accuracy and predictive validity. Psychological Bulletin, 136, 1092–1122.
  18. Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80, 1011–1027.
  19. Murphy, A. H., & Winkler, R. L. (1977). Reliability of subjective probability forecasts of precipitation and temperature. Journal of the Royal Statistical Society, Series C, 26, 41–47.
  20. Lichtenstein, S., & Fischhoff, B. (1980). Training for calibration. Organizational Behavior and Human Performance, 26, 149–171.
  21. Mellers, B., et al. (2014). Psychological strategies for winning a geopolitical forecasting tournament. Psychological Science, 25, 1106–1115.
  22. Chang, W., Chen, E., Mellers, B., & Tetlock, P. (2016). Developing expert political judgment: The impact of training and practice on judgmental accuracy in geopolitical forecasting tournaments. Judgment and Decision Making, 11, 509–526.
  23. Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of Personality and Social Psychology, 54, 569–579.
  24. Kross, E., et al. (2014). Self-talk as a regulatory mechanism: How you do it matters. Journal of Personality and Social Psychology, 106, 304–324.
  25. Grossmann, I., & Kross, E. (2014). Exploring Solomon’s paradox: Self-distancing eliminates the self–other asymmetry in wise reasoning about close relationships in younger and older adults. Psychological Science, 25, 1571–1580.
  26. Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119, 254–284.
  27. Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77, 81–112.
  28. Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119.
  29. Hudson, N. W., Briley, D. A., Chopik, W. J., & Derringer, J. (2019). You have to follow through: Attaining behavioral change goals predicts volitional personality change. Journal of Personality and Social Psychology, 117, 839–857.
  30. Ng, J. Y. Y., et al. (2012). Self-determination theory applied to health contexts: A meta-analysis. Perspectives on Psychological Science, 7, 325–340.
  31. Silverman, J., & Barasch, A. (2023). On or off track: How (broken) streaks affect consumer decisions. Journal of Consumer Research, 49, 1095–1117.
  32. Breines, J. G., & Chen, S. (2012). Self-compassion increases self-improvement motivation. Personality and Social Psychology Bulletin, 38, 1133–1143.
  33. Hadot, P. (1998). The Inner Citadel: The Meditations of Marcus Aurelius. Harvard University Press.
  34. Oettingen, G. (2014). Rethinking Positive Thinking: Inside the New Science of Motivation. Current.
  35. Ellis, A. (1962). Reason and Emotion in Psychotherapy. Lyle Stuart.
  36. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive Therapy of Depression. Guilford Press.
  37. Hofmann, S. G., Asnaani, A., Vonk, I. J. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36, 427–440.
  38. Fischhoff, B. (2025). Fifty years of hindsight bias research—Reflection on Fischhoff (1975). Journal of Experimental Psychology: Human Perception and Performance, 51, 143–150.
  39. Jain, S., Park, C., Viana, M., Wilson, A., & Calacci, D. (2026). Interaction context often increases sycophancy in LLMs. Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems, 1–26.
  40. Hong, J., Byun, G., Kim, S., & Shu, K. (2025). Measuring sycophancy of language models in multi-turn dialogues. Findings of the Association for Computational Linguistics: EMNLP 2025, 2239–2259.
  41. Thielmann, I., et al. (2026). A Consensus Statement on self-knowledge conceptualization, measurement, outcomes and changeability. Nature Reviews Psychology, 5, 338–351.
  42. Karimi Motahhar, V., & Gruca, T. S. (2025). How does training improve individual forecasts? Modeling differences in compensatory and non-compensatory biases in geopolitical forecasts. International Journal of Forecasting, 41, 487–498.
  43. Cui, Q., & Ishii, K. (2025). Research on self-distancing: A review from a cultural psychology perspective. The Japanese Journal of Experimental Social Psychology, 64, 84–101.
  44. Fyfe, E. R., Grenell, A., & Vest, N. A. (2026). The negative effects of self-focused feedback on mathematics problem solving. Applied Cognitive Psychology, 40, e70239.
  45. Kim, M. M., et al. (2026). Breaking negative cycles: A reflection-to-action system for adaptive change. Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems, 1–22.
  46. Sheeran, P., Listrom, O., & Gollwitzer, P. M. (2024). The when and how of planning: Meta-analysis of the scope and components of implementation intentions in 642 tests. European Review of Social Psychology, 36, 162–194.
  47. Haehner, P., Wright, A. J., & Bleidorn, W. (2024). A systematic review of volitional personality change research. Communications Psychology, 2, 115.
  48. Sezer, B., et al. (2025). Goal motives, approach/avoidance appraisals, psychological needs, and well-being: A systematic review and meta-analysis. Motivation Science, 11, 259–276.
  49. Lee, K., de Vries, R. E., & Ashton, M. C. (2024). Self/observer agreement in personality assessment by observers’ relationship types. Journal of Research in Personality, 113, 104529.
  50. Martin, M., & Mandel, D. R. (2025). Calibration feedback with the Practical scoring rule does not improve calibration of confidence. Futures & Foresight Science, 7, e199.
  51. Schertz, K. E., Orvell, A., Chandhok, S., Vickers, B. D., Moser, J. S., Ayduk, O., & Kross, E. (2025). The frequency, form, and function of self-talk in everyday life. Scientific Reports, 15, 38883.

Titles are given in English, as published.

Journals keep the past. Advisors offer opinions.
Avrelius keeps the evidence — something firmer than memory and mood to rely on at your next crossroads.
Join the waitlist →← Back to the main page