Publications by Bill Cornelius
--- Sitemap ---

astrology statistics
Freeware Download (mac only)
508K, Beta v. System 9 & earlier.

Horostat 1.0 ReadMe.

The Horostat software application offers the statistics to either prove or disprove Astrology.

Table of Contents

Horostat blog
Possible proof positive
What is it?
How to use it
What's it for?
Recognizing garbage
Functions & buttons
Entering chart data
Changeing Aspect orb width
Elapseing a chart
Click 'n Drag planets
Aspect line ID
The Statistical part
Selecting data
List of known bugs

What is it?

In the 1970's, the Dept. of Paranormal Studies at Stanford Univ. did a statistical analysis of astrological interpretations, also a more recient one 1 by the Univ of Toronto, their result was that there is no special correlation between planets and events. I believe their result is false, because though the Stanford Univ. data may be correct, their interpretation of "event similarity" (my quotes) was too narrow, based as it were, on traditional (medieval) concepts of natural philosophy rather than science, it only proves that one of many interpretations were unfounded but shouldn't be extended to Astrology as a whole. So I wrote the Horostat program to provide a sketch of any verifiable influences and I think I found at least one that I've never heard of as well as some supporting evidence of more traditional ones. See Horostat Blog, 24 Jan. 2011.

Q: So how does a person recognize a "similar" event?
A: One trusts certain memories which are obtained through experience rather than through communication (books & lectures).

Simply put: if it's not yours, you won't recognize it.
Less Simply put: Awareness which is not socially relevant is subjective by default, though it occupies the greater usage of ones mental CPU. 2 Social awareness is built on a learned verbal description of what's important, which limits what we allow ourselves to know to mostly things which we can verbally communicate. Those sources (social conventions) require trusting someone else's experience over our own, and though that process is the backbone of both fundamentalism and primate society, those sources may still be in error. (An irrefutable example is the process that elected George Bush II. Twice!). Other examples: Historical Figures in literature (wikipedia)

That's not saying social conventions are false, we depend on them for group cohesion. Without them what would music or humor be like? But since no two people remember the same event the same way, any attempt to define astrological events is equally open to disagreement. The purpose of the Horostat program is to give statistical evidence of astrological influence through ones own subjective recognition of "similar" events. It's the programmers belief that recognition of event similarity is an emotional memory, like remembering dreams. However: objective, verifiable, "hard" information depends on another persons agreement. We look to other people to see what's true 3, so some information is socially recognized as being an event and some ignored 4. The value of information is in its usefulness to the people in communication. Everyone sees UFOs, some people see the Blessed Virgin, no one sees J. Edgar Hoover.
So: in order to appraise information that MAY be true but can't be proved without the dilution of another persons opinion of an event that they didn't actually experience, Horostat's statistical function associates subjectively perceived events with verifiable astronomical events and give odds to one against random chance. The results don't lie, but what they represent can, so Horostat doesn't attempt to make sense of them: ie. no interpretation printout. (also because there are already hundreds of Astrology books and programs describing other peoples opinions).

Various disclaimers:

There are a number of ways to draw conclusions from statistical data. This program uses a method called "frequency probability", but some say a technique called "Bayesian probability" is closer to the way a human brain actually works. 5 6

The writer/programmer (me) admits to having read one (1) book on statistics as a template to model this software, so any faulty procedure is from ignorance and not from intent. If someone want's to do a better job, I'll be happy to send you a copy of the source code. As of Oct 2012 I've given away 4 copies (and haven't heard any bad reports). 7.

Horostats House System is 30 degree cusps with the planets all aligned exactly in the middle of the Zodiac band. This way, everyone is equally dissatisfied. (I don't like it much either, but only because no one agrees). Horostat's houses definitions are based on constellation groups (Signs), and not on specific associations with fixed stars.

How to use it:

To use the statistical function, one needs to categorize dates by event such as from a diary, postmarks on response mail from mail order business, insurance claims, police or hospital entry records, etc. Mail order responses may be classified by product or volume for a profile of what inspires a response to your ad (it should cover several publication dates) Also works for hits on a web site. Hospital records might be by volume or affliction. There are 2 kinds of errors possible in selecting material:

  • The first is like shooting arrows at a barn and drawing a target circle around the best group. This is called a Post-defined hypothesis, which is a blasphemy against mathematics. Statisticians consider people who do that to be capable of pedophilia.
  • The other is Definition Creep in subjective indexes. In a dream diary, a person can make indexes of anything describable, but these kinds of subjective events open a door to faulty procedure, not when the information is read, because literature does tend to recall mostly consistent memories 8 9, but rather when the information is recorded. This is because the definition of those events is bound to undergo subtle changes in interpretation from day to day, as ones emotional state is bound to fluctuate (like from winning or loosing the Lottery or OMG just anything), . . planets y'know.
      The writers solution to "Definition Creep", and an Undefined Hypothesis, is to write down exactly what you expect to find. This is called a Predefined Hypothesis. This written thing is: The Category Definition, it identifies the goal, and it's a big deal. The program prompts you for a Category Definition in each folder, every time you open a new blank text window, till finally you save one for each folder. In it, you describe what the files in the folder have in common, what you are looking for, and that description is included with any statistical analysis done for those files. Don't cheat or you'll destroy your reputation before you even have one. A post defined Hypothesis isn't the same thing as looking for the effect of an astrologically induced event which can be verified objectively, and Horostat has no problem with that research. Ethically though you shouldn't cull the charts that don't agree and call the results authentic. (without first joining the Republican party)

    The following is an example of a category definition:

    The charts in this folder represent the influences on the result of decisive air battles in the Pacific theater during ww2. Specifically the influence of Mercury, Venus, Mars, Neptune and the Ascendant on the relative signs associated with America (Mercury/Gemini) and Japan (Venus/Libra). ...etc. etc.
    The chart examples also lists the time/date of influential events of those battles, which can be charted and added to the analysis. Such as times of sightings, engagements, maneuvers, damages, withdrawals, etc.

    The user must provide at least 5 dates-charts-files (they're all the same) per category to enter the Statistical function's graph box, and for increased accuracy, up to 500 per category, (though some say this is minimal for real evidence). With less than 17 dates, the results will be distorted. In the Stats. Dialog Box: the slide button labeled "<--real-odds-not-->" adjusts the stringency of the analysis by percent from 100% (slide to the left) to 0% (slide to the right, completely bogus). Lower percentages multiply a small sample till it's big enough to show which planets have potential to become players, given enough charts and a continuing trend.

    Ok, say I decide to find out if Lulu really loves me, I check my diary to find when Lulu occurs, but what I really find are those times when I was feeling the impulse to write down whatever the event was. Objectively, this may not be what I'm looking for. That's not saying "Whenever I Feel Like Writing It Down" isn't a justifiable category because, like with Schrodinger's cat, Quantum physics says the observer can't help but affect the result. Its just that statistics keepers take a straight faced vow to be objective (ie: it's best to get paid before any work is started. See above paragraph 3).

    The accepted scientific procedure is to write down EVERY contact with Lulu (or The Boogie Man), separated into categories represented by some sort of equally spaced grid of time or space. To establish this impartiality, Horostat uses the predictability of celestial mechanics as the grid. However, only the most outstanding events are selected because they represent the topic of the hypothesis, and because this is the way awareness works, as opposed to recording every detail of every event, which is time consuming and impossible any way because we get back to "what is an event?" (see ppg 1). ...Well, an event is a subjective perception, so trust yourself.

    Each event of a category can be rated by "perceived intensity". Of particular value are apparently spontaneous events. But just because I see Lulu every day by the bus stop doesn't mean its unimportant, it just means those might be rated with numbers closer to zero. One creates separate categories for different kinds of events, but one can rate the same event in different categories if they apply. One names the category title when the folder is created, and it's possible to rate the "intensity", by giving the first letter of the file name, a value from - 9 to + 9, (ie: "-1 Hairwashing " in the "Strife" folder). Rating the intensity in the title is for visual selection only, like an analog sub-category. The name does nothing to affect statistical analysis. Once a group has been selected they're all equal. But if results are not obtained in a narrow definition of the event, the scope can be widened by including those file names with numbers closer to zero in of the "perceived intensity" index and vise versa.

    Now, even with all this stuff spelled out, one still can't tell about lovely Lulu. So one decides to revise the category hypothesis to read: "How can I optimize my relationship with she of the beauteous derriere?"
    a) One enters all the dates with the highest numbers to find what motivates us (one).
    b) One enters all the dates with the lowest numbers to find when to avoid her.

    Another note on Definition Creep; When one asks Lulu to the Grand Ball, She says she has to wash her hair (for 48 hours). How does one classify this kind of event? A wrong interpretation will wreck a small sample of events, and skewer ones pining heart. So, if no verification can be obtained, one rates the event closer to zero and denies whatever one can afford (One truely believes she had cement in her hair), or create a new category.

    And last: LuLu decides to sell the skin business and move back to Suburbia so, should one date an event when one hears of it, or when the decision concerning the event was made, or when the event actually occurs? The answer is yes and put copies of all three in their own category. There may be no correlation but we'll never know if we don't look.

    My Dragonfly Wind Generator Company (buy one) seems to have a consistent lump of orders when the Sun enters a fire sign, so we figured this is impulse buying of a $800 item, and refined the ads to appeal to impulse buyers. But wait! a study of the dates of email orders indicated more sales were made during the winter. These signs are generally serious and not too impulsive, but it is the time of year when wind does it's thing best, and a serious person would notice the potential. See Horostat Semi-Demo.
    To REFINE the discovery of whatever is influencing a category of events, look thru the aspects (etc,) shown in the statistics graphs, then look through the charts in that folder for only those charts that have that particular aspect. See if the chart's notes have anything in common, and after wailing briefly about faulty procedure: select only those charts to see what else is happening there. The "Notes" button provides an opportunity to record anything outstanding in an individual chart, so those notes can later be reviewed for similarities.

    What to do with it:

    The design philosophy of this program is to provide a tool for general astrological research which will allow one to:

    (1) Identify the common influences affecting 5 or more dates.
    (a) Identify dates when those influences will be strong or weak.
    (b) Identify buyer / seller profiles.
    (c) Allow lead time using customer databases to anticipate trends.
    (d) Estimate the probable areas of harmony or disharmony among a group of individuals.
    (2) Obtain statistical evidence for an astrological concept.
    (3) Chart specific conjunctions, aspects or planet returns, and identify cycles.
    (4) Get filthy rich by dabbling in the stock market.
    (5) Provide statistical proof that one is not mad. (heh ...)

    Identifying Garbage:

    Any group of samples will have something in common, the question is whether it's relevant to your Predefined Hypothesis. If the results are consistent throughout a series of samples of varying sizes, and missing in samples that don't represent the Predefined Hypothesis, chances are it's relevant. But even then, just because some odds turn up doesn't mean they're proof of anything. For example:

    • I want to find out which influences affect the Santa Anna winds, I put in some dates and viola! Uranus, Neptune & Pluto are in yellow, BUT the sample covered a shorter span of time than their transit through one sign, so all charts would have them in the same sign and they would show high odds for being there.
    • The Sun, Mercury and Venus are usually in each others 1st, 2nd or 12th house anyway because Venus is never more than 48 degrees from the Sun, and Mercury is less.
    • The Midheaven and Nadir are always 180 degrees apart so they're always in each others 7th house.
    • The Suns 4th and 10th house for Ascendant and Descendant.
    • If enough charts are drawn excluding the calculations for the horizon (because the time is unknown, ie. zero in the hour box, see next section), it affects the Part of Fortune and puts it in whichever house is as far behind the Moon as the Sun is behind the Ascendant, which is about 3 signs at noon.
    • There are probably others.
    A fair rule of thumb of what to watch for might be: Warmer colors will naturally fall to the slower planets, so if any of those are the cooler colors, it may bear investigation, the same goes for warmer colors among the faster planets. A lot of hits in a column probably means something. But for me, too many are confusing so I concentrate on the ones high in the column and take aspects between the sun mercury and venus with a grain of salt because they form aspects with each other all the time and some are bound to show up in any series. However, those very random aspects may be a primary instigation for people to act on any whim that they'd otherwise sleep through. And if NONE show there, that's noteworthy.

    The Horostat functions:

    1) Entering chart data: access the first window or by clicking on "New Chart" or key: (Cmd. "c"). To select another data editing box, click on it, hit tab, or key: (Cmd. "1" = day, Cmd. "2" = month, Cmd. "3" = year, etc).

    If you're doubtful about the hour of an event, enter zero (0) in the hour box before saving. This "disconnects" the hour box for that chart, and prevents unreliable entries from swamping the data associated with the horizon. (the chart will be drawn for noon, however, this goofs up the calculations for the Part of Fortune (if you use the Arabic Parts) and reduces the reliability of it's figures) If you know the hour but not the longitude, click the box labeled "Approx. Asc." and the horizon will be placed according to international time zones (though they are seldom consistent with longitude, continental US is within 3 hrs, i.e. +/- 1.5 hrs).

    Data entry is a bottleneck, it's ok to enter few events at a time this way, but listing 300 dates is grueling, & tedious beyond belief. So the program has a function in the text editing window (see below 'Notes'). A menu called 'Import Data' instructs the program to read from a text file and record hundreds of charts in a couple of seconds. They still have to be entered but it cuts the time drastically. The text has to be in the format below. A word processor program with spreadsheets can be set up to list the info. but if you haven't used a spreadsheet before, it won't save time. Instead use a regular textfile w/ a fixed width font like courier and justify everything to the right. That way goes pretty fast and errors stand out because the only variable sized entry is to the left. And if you paste this string: " AD, Lat.--N, Long.---W, ", it flies. For the name there should be no caps and total less than 21 characters, + one space at the end of the name.

    name_id_for_the_date , 21/04/01952 AD, Lat.40N, Long.119E, 10:30 PM
    another name , 02/04/02002 AD, Lat.39N, Long.123W, 12:46 AM
    name 3 , 15/09/02002 AD, Lat.--N, Long.---W, 00:00 AM
    name_4 , 30/11/01999 AD, Lat.33N, Long.101W, 03:31 PM

    + 246 more

    The format is: name , dd/mm/yyyyy AD,(or BC) Lat.LLN(S), Long.LLLW(E), hh:mm AM (or PM). If you don't know the Latitude and longitude, replace those numbers with dashes (-). If the time is unknown, replace the hrs & mins w/ zeros. Punctuation characters, spaces, and the number of characters after the name are all critical, a total of 47 characters or spaces, not including the name.

    Before these files can be used in the statistic function, they need to be compiled using the "Rebuild" button in the statistics dialog box.

    2) Precision: Charts may be calculated for any moment from 17700 BC to 17700 AD,
    though with some suspicion of accuracy since verifying records are scarce before 1600. With the model used, predictable error will be less than 1 second of one degree, within 100 years of 1900. So, with that latitude for error, one may expect a linear accuracy of less than one half degree within 17700 years of 1900, (not including the Moon or Pluto).

    3) Notes may be recorded in fixed association to a chart by opening a file (automatically created with each chart saved) through the menu bar, or Cmd."o" when any chart is already open. Re- saving a previously saved chart will erase it's notes, but they may be selected, copied and pasted to a new charts notes. Independent textfiles may also be written and saved, opened with key Cmd. "n", ('new' or 'note').

    4) Aspect orbs can be adjusted or deleted through the Aspect box from the chart page thru it's own button, or key: (Cmd. "a"). To delete a single aspect from being shown, set it's slide button to "off". To delete a group of aspects, such as Kepler's aspects which some believe to be bogus, or because it pleases your fickle nature, click on any of the numbers that indicate the orb width. This darkens the row of aspects and shows that they are off. Same for the "Delete" button to turn off/hide all aspect calculation. "Cancel" cancels any changes.

    5) A chart may be elapsed by any amount of time up to +/- 4000 years to observe the horizon or a planets movement. To access the Elapse box, key (Cmd. "e"). This function doesn't have much to do with statistics but it's interesting to watch planets interact in motion. For example: the outer, more "malefic" planets move slower than the inner ones, as the faster planets overtake slower ones, and they exchange 1st and 12th house position, so one can observe the process of short term solutions producing long term effects. Hold down one of the arrow keys. Use the Elapse function also if you know the time between events and one of the dates.

    The spring equinox is always to the left to give a consistent point of reference while observing planetary movement. Planet returns have their own buttons.
    6) Click 'n Drag a planet to a new location to find out WHEN a planet will be there, a date will be compiled for approximately the middle of that degree on the next or last time it was there. To find an EXACT aspect (▒ a few minutes of degree), press the apple command key while Click/dragging the planet. It's only possible to drag the faster planet to the slower ones, a successful combination is indicated with a green circle on the degree of the aspect. A red circle means you have the slower planet, so drag the other one. To find out when a planet rises or sets: drag the descendant to the planet, or set the elapse to 4 minutes and hold down the arrow key. A fast computer makes a sort of animated movie of the planets movements. For slow computers press: Cmd. F. That deletes the aspect calculations so it goes a bit faster.
    7) "Help" is available almost everywhere by clicking on a Help button or key: (Cmd- H). Help describes what functions and key commands are available in any dialog box.

    8) If there's a confusion of aspect lines in the chart, click on any in question to find out what they are.

    9) The statistical function:

    determines which astrological influences are common to a series of charts, then gives odds against chance.

    Get there from the chart window by clicking the "Statistics" button or key (Cmd. "g"), select the files you wish to use with the interface box shown below. Click on the list of files that appear in the left column, they move to the right column where they can be alphabetically listed by name, time, day, month, year, latitude, & longitude. De-select any that don't fit your profile.

    Data Selection Window

    It may be necessary to standardize the aspect orbs of the selected charts IF the Aspect Box settings were changed at any time during the span of time that the charts were created. There's a button for this, which eliminates orb variations that may give erratic results. Usefully named "Rebuild", it opens the Aspect Box (the Aspect Box alone is also accessed from the chart page thru it's own "Aspects" button, or key: (Cmd. "a")), so you can verify that they are the way you want, then it goes through the (above) list on the right & adjusts them all to those settings. This button is also used to build a folder of new imported files (see Data Entry). In this process, if any files are found to be corrupt or there by mistake, they're removed from the right hand column, and their names are listed, but they are left unchanged in the category definition folder.

    Then click the "Graphs" button or key (Cmd. "y"). The window below will pop up (without the colored boxes at first). To begin, select one of the boxes outlined in red which represents the type of analysis you wish to observe: Positive odds, negative odds, or any odds (both positive and negative combined). Then a box outlined in blue. The resulting statistical ranking will place the highest odds (but not necessarily the most significant) on the top line and the lowest on the bottom. The window below shows positive odds of planet dispersion through the 12 houses starting from the first degree of whatever sign the planet was in (cHouse, means classic house system, xHouse means generic house system where the slower planet represents zero degrees of the first house), for 149 charts in a category folder named A Trial. The columns of planet symbols beside a number show the highest occurrence of the planet in the house whose number is shown. For example: odds that Mercury was in the 11th house of Mars (top row, 6th col.), are up to 9.92 to 1. because it's light blue, Those colored boxes on the left range groups of odds by color to give generalized visual context, but the group divisions are arbitrary for rough appraisal only. Exact odds are below.

    To view the exact odds of any individual box, simultaneously press Command and click the mouse, like this below where we see the odds for Mercury in the 11th house of Mars was actually 5.87 to 1. The list on the left shows the highest odds in that color category. The colors make differences jump out but are just for rough appraisal, they generalize.

    Below are the odds for any planet being in any sign for a Category folder named "DOW falls", with 34 chart files. (Full disclosure: it's just an example. Pluto was last in Taurus in 1882 and the DOW was created in 1896.)
    The Report button ("Rprt" lower right above "Done") determines the next date when those influences will occur, progressed from the date shown in the "New Chart" dialog box. The report may appear unpractical because it can list the next occurrence of fast planets happening much sooner than the next occurrence of slow ones so that any correlation of their influences would be moot. For example: Your report says that significant aspects are: Venus and the Moon, and Saturn and Mars. Venus aspects the Moon on Jan 21, but Saturn won't aspect Mars till 16 months later. For the data to be relative, the influences should overlap the way they would in a significant number of the charts being analyzed, so go back to the "New Chart" dialog box and put in a date slightly before the next occurrence of the slowest planet of interest, then redo the same statistical analysis. The report will then be done for the next occurrences of those planets which can (but may not necessarily) overlap the occurrences of the slowest.

    Planet search accuracy is rounded to the degree for the inner planets and up to 4 decimal places for the outer ones, for example: the horizon passes through 1 degree in four minutes, but Pluto may take years, so more decimal places increases Plutos search accuracy to within days. When searching for the beginning or end of an aspect, or a planet in a given sign, the search may stop when the planet leaves the aspect or sign even though the planet may return shortly while the aspects are still in influence. (it happens). And it shows no indication, so be aware that although this program does reveal many possibilities, you may confidently indulge suspicions that it does not hit them all.

    • Horostat's ephemeris is from the book: Astronomical Formulae for Calculators, fourth edition, by Jean Meeus, published by Willmann-Bell inc. 1988
    • The statistics are from: A Primer of Statistics For Non-Statisticians, by Abraham N. Franzblau, published by Harcourt, Brace & World, Inc. 1958.
    • Programming advised by Larry Martin, Mendocino ROP. Using Think Pascal 4.0. on a 68k mac and Mac plus (@ a blazeing 15 hertz, Yow!!).
    • Horostat was written and rewritten and rewritten and (rewritten)10 by Bill Cornelius, PO Box 57, Albion CA 95410
    List of known bugs:
    1) The main bug is the internal search engine. It seems reasonable to provide a way to check on any statistical result by finding the next time to expect the event type in question. However I must apologize for loosing patience with the code involved, I can see from my website statistics that the Horostat program is downloaded sometimes 5 times a week, so somebody is interested. But as of Nov 2004 I've received only 2 e-mails containing feedback, and then because I started requiring it before downloading. If you find any bug to be a problem, or that the searches are critical, let me know and I could be convinced to wade into it again, (with enough flattery or bribery). It's effect is usually obvious, (for example, the last entry of Venus or Mercury into a sign will sometimes search in unlikely places.) Actually I have a lot of really good excuses why the dang thing won't work reliably, but listing them just sounds lame, so instead I condescendingly to refer to them as technical aberrations. If you're a geek, and have to know, then click here. OK I fixed some of that stuff, are you happy yet? If not, send money & I'll think about it.
    2) The "Print" command is nonfunctional. However, the text files may be opened and printed from most word processor applications. and the charts can be printed from a screen snapshot, Mac: (cmd-shift-3), or (cmd-shift-4).
    3) The Latitude is fixed at 39 North.
    4) The window identifying aspects sometimes mistakenly prints "applying" for "departing" and vise versa, except for the north node.
    5) Some of Msr. Meeus results disagree with those of the US Naval Observatory, but seldom by more than a few minutes.
    6) Some people say Libra is on the wrong side. I don't apologize for that.
    7) There are tiny harassments resulting from slapdash programming which blocks-in the main functions and doesn't return to clean them up: some mouse-clicks work erratically, stack overflow crashes, print appearing in the wrong place. None are fatal and most can be ignored or worked around. Don't complain, it's free, and an Aries did this (& suprise! it's not finished) so get over it.

    A treatise touching the Defence of Astrologie A digital edition of Simon Forman's & Richard Napier's medical records 1596-1634. more
    +/- 30 Astrology software programs. Hungarian website w/ translator. Software is mostly in English.
    50+ Shareware astrology programs Italian website, the programs are in several languages.
    Aspects between planets and their influence over financial markets. Dharmic market timing for Traders
    A comprehensive directory of Astrology and prediction resources.
    Joe's AstroCalculator v 1.32 + software
    XEphem: The Astronomy Software for UNIX and Linux
    JPL Solar System Dynamics
    Astrology HQ: categorized resource directory for everything about astrology.
    Astrology Weekly

    If you get good results, or feel something should be included or changed, please send any short rants to If possible, please include your category definition &/or a copy of your report.

    ęBill Cornelius, 15/Jan./1997