On My Astrology Records and Preliminary Statistics: Several Questions I Want to Face Honestly
Through several recent discussions, I have become increasingly aware that different people have very different ideas about what astrology is and what it is supposed to do.
Some people see astrology as a predictive technique and believe that it should be able to make concrete judgments about future events before the outcome is known. Others see it as a symbolic language used to organize experience, express feelings, and understand life situations. Some place greater emphasis on its role in consultation, emotional support, and self-understanding. Others regard it as a traditional system of techniques, a spiritual practice, or a model for interpreting the world.
As a result, many debates that appear to be about whether astrology is “accurate” are not actually discussing the same question.
Before discussing whether astrology works, perhaps we first need to ask:
What exactly do we mean by “astrology”? What do we expect it to accomplish? What kind of result would count as successfully accomplishing that task?
An astrological method may be valuable on one level without necessarily being equally valid on another.
For example, astrology may be a rich and expressive symbolic language, but that does not automatically prove that it has predictive power. It may help people organize their emotions during consultations, but it may not necessarily provide information beyond ordinary psychological support. It may also produce correct judgments in certain cases, while still leaving us with further questions: Are those judgments consistently correct? Can they be repeated? Do they genuinely provide information beyond common sense?
1. What I Am Actually Trying to Investigate
For this reason, I no longer find it very useful to ask the broad question:
“Does astrology work?”
The question is too large, and it allows everyone to answer according to their own definition of what “working” means.
I would rather break it down into smaller questions:
Within a particular astrological method, and for a particular type of question, what does the method claim to accomplish? What is the standard of success? Is its value expressive, interpretive, consultative, or predictive? If it claims to have predictive value, can it make specific and testable judgments before the outcome is known? Does it provide additional information beyond common sense, base rates, or ordinary experience?
What I am trying to do may not simply be a study of how “accurate” astrology is. I may also be gradually constructing a map of astrology’s functions and boundaries: which functions appear relatively clear, which claims still require testing, and where subjective interpretation, memory bias, and retrospective rationalization may affect our conclusions.
2. What Data I Have Included So Far
I have recently reorganized part of my past astrology records.
It is important to clarify that my current statistical results come from only one limited batch within a much larger body of material. They do not represent all of my astrological cases.
At present, this batch includes only material that meets the following conditions.
2.1 Only My Own Records
The current analysis concerns astrological judgments that I personally made in the past. It does not include cases from other astrologers.
Therefore, these results can at most reflect my own recording habits, judgment methods, and practice. They cannot be directly generalized to other astrologers, and they certainly cannot represent astrology as a whole.
2.2 Only “Astrology-Only” Judgments
As far as possible, I have retained only material that can be clearly identified as an astrological judgment. I have tried to exclude conclusions that mainly came from background information, common-sense reasoning, other methods, or explanations added only after the outcome became known.
However, defining “astrology-only” in operational terms still requires further refinement.
In real consultations, it is often difficult to separate astrology completely from common sense, communication experience, and knowledge of the client’s circumstances. This is one of the problems that future statistical work will need to address.
2.3 Clustering Was Reduced
Within a single question, there may be several closely related judgments. The same event may also be recorded repeatedly or discussed across several messages.
If all of these highly correlated statements are treated as independent predictions, the same underlying event may be counted multiple times, making the sample size appear larger than it really is.
For that reason, I carried out an initial process of reducing clusters, or “declustering,” in order to avoid repeatedly counting the same event or the same group of closely related judgments.
However, deciding what counts as one independent case, and determining whether two predictions belong to the same cluster, still involves some judgment. Clearer and more consistent rules will be needed in the future.
2.4 Only Astrological Judgments Explicitly Written in the Text
I included only judgments that were actually written down at the time. I did not rely on retrospective memory, and I did not count thoughts that I merely remember having had but did not record.
This is because memory is easily influenced by later outcomes. After an event occurs, we may remember the parts that resembled the result more clearly, while forgetting what was incorrect, unclear, or never actually stated.
For me, only written records created before the outcome occurred are reasonably suitable for later analysis.
3. Why I Have Not Yet Included Client Cases
The current batch does not include astrology cases involving clients.
The main reason is privacy.
Client cases may contain sensitive information about personal experiences, family relationships, health, work, romantic relationships, or finances. Even if names are removed, deleting names alone may not be enough to protect identity. A combination of life details may still make a person identifiable to themselves or to people who know them.
Therefore, if I analyze this material in the future, I will first need to carry out a more complete anonymization process. This would include removing or modifying identifiable details, rewriting event descriptions, and carefully considering the boundary between what is statistically necessary and what is private.
Compared with organizing my own public or private records, this would be a much larger project.
For that reason, I have not yet begun this batch, and I do not intend to lower my privacy standards simply in order to obtain more data more quickly.
4. My Preliminary Observation So Far
Based on the limited material I have currently organized, together with my reading of related discussions and statistics provided by others, my present view is that:
Astrology may have a stronger function in interpretation, organizing experience, and providing companionship than it does in prediction.
However, I want to emphasize that this is only a preliminary observation, not a final conclusion.
First, my earlier recording habits were not designed for rigorous statistical analysis. Many statements were not written in a falsifiable form. In other words, they may not have been specific enough to allow a clear distinction between “correct,” “incorrect,” and “vaguely consistent” afterward.
For example, if a judgment is very broad, uses highly abstract language, or permits many different retrospective interpretations, then even if it later feels as though it “came true,” it remains difficult to determine how much genuine predictive information it actually contained.
Second, after reading more of the existing research on astrology, I found that several influential experiments did employ relatively rigorous research designs. These included double-blind procedures, criteria and methods of analysis established in advance, and, in some cases, the participation or approval of astrologers in determining the testing conditions. Some studies also provided sufficiently detailed descriptions of their procedures to allow later examination, discussion, and reanalysis.
I therefore do not want to question the rigor or fairness of this research simply because I have not personally reviewed every original dataset in full. A substantial body of controlled research already exists, and its overall findings have generally not shown that astrologers can perform consistently, replicably, and above chance under blinded conditions. These results deserve to be taken seriously and form an important part of the background against which I consider claims about astrological prediction.
At the same time, I do not want to overextend these findings into a complete rejection of every form of astrology, every astrological technique, or every claim associated with it. Each study examines particular methods, tasks, samples, and definitions of success. The design and statistical interpretation of some well-known experiments have also remained subjects of later discussion.
My current understanding can therefore be stated cautiously as follows:
Taken as a whole, the more rigorous empirical research conducted so far has not provided sufficient evidence that astrology possesses a stable, reliable, and replicable objective predictive ability beyond chance, common sense, or information already available.
This does not mean that every related question is no longer worth investigating. It means that any claim that astrology has objective predictive ability must engage seriously with the existing findings and be supported by evidence that is clear, testable, and reproducible.
My analysis of my own records is not intended to overturn or dismiss the research that already exists. I am asking a more personal and limited question: when I examine my own practice using stricter and more falsifiable methods, will I obtain similar results? - How many apparently accurate judgments actually fail to provide information beyond common sense or base rates? And might astrology’s more consistent value lie primarily in interpretation, the organization of experience, and companionship rather than in objective prediction?
5. Does Prediction Provide Information Beyond Common Sense?
Building on the previous point, another question must be considered: even when an astrological prediction appears to be correct, does it provide information beyond common sense, base rates, or what was already known?
In addition to asking whether a prediction eventually turned out to be correct, I believe there is another question that must be addressed:
Does the astrological judgment provide incremental value beyond common sense, base rates, and information already available?
Some events already have a high probability of occurring.
If something was very likely to happen even without astrology, then predicting that it would happen does not automatically demonstrate that astrology contributed additional information.
Conversely, a judgment with genuine predictive value would ideally meet several conditions:
It is stated before the outcome is known.
It is specific enough to be judged correct or incorrect.
It does not depend on changing the interpretation afterward.
It does not merely predict something that was already highly probable.
It genuinely adds useful information beyond common sense or the base rate.
For this reason, in the future I do not want to record only how many predictions were right or wrong. I also hope to compare astrological judgments with common-sense expectations or estimated base rates.
This will not be easy, because “what common sense would predict” also needs a reasonably stable standard. It cannot simply be decided retrospectively, after the outcome is already known.
6. How I Changed My Recording Method This Year
Beginning this year, I have tried to record astrological judgments in a more rigorous and falsifiable form.
For me, “falsifiable” means at least the following:
The judgment is written down before the outcome occurs. The subject, content, conditions, and time range are stated as clearly as possible. The criteria for what would count as correct or incorrect are specified in advance. After the outcome is known, the original interpretation is not casually widened or altered.
At present, I do not yet have enough new records to produce a formal statistical result.
I also do not believe that simply using the word “falsifiable” automatically makes a record rigorous. Questions such as how to treat partial accuracy, timing errors, multiple simultaneous judgments, correlated events, and cases whose outcomes cannot be confirmed all require relatively stable rules established in advance.
Those rules may continue to change as the work develops, but any changes should be clearly documented rather than introduced in order to make earlier results look better.
7. I Am Not Publishing This to Prove Anything
I want to state this especially clearly.
I am not publishing this material in order to prove that astrology definitely has predictive power. Nor am I trying to prove that I personally am an accurate astrologer.
The current data are not sufficient to support either conclusion, and I do not want to present an ongoing process of self-examination as though it were already a proof.
The main subject of this experiment is not the public. It is myself.
Above all, I need to be honest with myself.
If a judgment was vague, I want to acknowledge that it was vague. If it did not provide anything beyond common sense, I want to take that into account. If it was wrong, I want to preserve the failure rather than display only successful cases. If future statistics do not support some of the claims I previously believed, I hope I will be willing to narrow, revise, or even abandon those claims.
Likewise, if future data genuinely indicate that some astrological methods have stable predictive value under particular conditions and provide information beyond common sense, I hope to examine that carefully and transparently rather than relying only on a few impressive cases.
For me, facing the results honestly is more important than defending a predetermined position.
8. I Am Looking for People With Similar Recording Practices
There is also a practical reason why I am publishing this article.
I would like to know whether anyone else is carrying out similar forms of recording and testing.
For example:
Does anyone write down specific astrological judgments before the outcome is known, in a form that can later be judged correct or incorrect?
Does anyone preserve failed, vague, and unverifiable cases rather than collecting only successful examples?
Has anyone attempted to deal with duplicate counting, case clustering, and sample selection?
Has anyone compared astrological judgments with common sense, base rates, or non-astrological judgments?
Has anyone developed a relatively complete anonymization process for protecting consultation clients?
Is anyone willing to discuss their criteria, failed cases, and methodological limitations openly?
I am not necessarily looking for people who share exactly the same position as I do.
We may disagree about what astrology is, whether it has predictive ability, and how it should be studied statistically. What matters more is whether we are willing to state our standards clearly, preserve records that can be tested, and face results that do not match our expectations.
If anyone is doing similar work, I would be very interested in exchanging methods, difficulties, experiences, and reflections in the future.
We do not need to work together in order to prove a particular conclusion. We may instead help one another identify methodological blind spots, compare different classification systems, and discuss which questions remain unresolved.
9. Disclosure Regarding AI Assistance
Because I am frequently asked about the textual source of this article, I would also like to disclose the following openly.
The central ideas, viewpoints, research direction, statistical scope, and personal position expressed in this article are my own.
AI assisted me in reorganizing the structure, refining the language, and translating parts of my Chinese writing into English, especially when converting my original Chinese thoughts into clearer English expression.
Afterward, I personally reviewed and revised the article.
AI participated in the organization of the text, but it does not take responsibility for my views, nor is it the subject of the research or judgment presented here.
I take full responsibility for the statements, methodology, interpretations, and any possible errors in this article.
In my view, using AI to assist with language does not reduce the author’s responsibility. Regardless of what tools are used in the writing process, the person who decides what to publish, what to claim, and what to stand behind remains the author.
Conclusion
After these recent discussions, I increasingly feel that astrology may not be a question that can be answered only by choosing between “belief” and “disbelief.”
Different people may be using the same word while engaging in very different activities.
Some seek prediction. Some seek understanding. Some seek companionship. Some study traditional techniques. Others treat astrology as a symbolic language.
Therefore, before evaluating astrology, perhaps we first need to clarify:
What do we believe astrology is? What do we believe it can do? What standards do we use to judge whether it has succeeded? What kind of result would cause us to reduce our confidence in a particular astrological claim?
My own position remains open.
Astrology can be a meaningful symbolic language, but meaning at the level of language does not automatically demonstrate predictive ability.
Whether astrology can predict, what it may be able to predict, and under what conditions it might outperform common sense must all be tested separately.
Future results may support some claims. They may also force me to narrow, revise, or abandon others.
Whatever the outcome, I hope I can at least do one thing:
Not try to prove astrology, and not try to disprove astrology, but observe as honestly as possible what I actually see.