The Probability of the Capstone The Capstone probability analysis concerns the agreement of a fixed network of scriptural intervals with specified Hebrew calendar dates and a larger chronology terminating in 2035. Its numerical method counts permitted interval connections, evaluates the calendar requirements together, and joins those results under a stated probability law. The calculation preserves the source relationships, the counting convention of each interval, and the distinction between constructing the network and testing its calendar placement. The research began with the connections described in Scripture. The completed structure emerged as those connections accumulated. The original core preceded the recognition of its calendar overlay. The research entry dated May 23, 2013 records the earlier arrangement and its apparent festival correspondences. Subsequent study supplied further required intervals and clarified counting conventions. By 2018, the 2,520-day, 2,556-day and 430-day requirements had become part of the completed specification. The macro chronology was developed later. This sequence establishes the methodological order recorded in the research: textual derivation, calendar observation, further constraint, and comparison with the macro chronology. Daniel supplies the 2,300 evenings and mornings, the 1,290 days and the 1,335 days. Revelation supplies the 1,260 days and relates the wilderness period to time, times and half a time. The forty-two months appear alongside the witness period and the trampling of the holy city (Daniel 8:13-14; 12:7,11-12; Revelation 11:2-3; 12:6,14; 13:5). In the Capstone construction, the event descriptions identify common beginnings and endings. The removal of the sacrifices anchors the shorter measures, and the restoration of the sanctuary identifies the common final boundary of the 2,300 and 1,335 measures. These event identifications generate numerical relationships before a calendar is consulted. The 1,290 extends thirty numbered days beyond the 1,260. A shared terminus for the 2,300 and 1,335 places their beginnings 965 days apart under the elapsed conventions used for those two measures: 1290-1260=30, 2300-1335=965. The calendar comparison therefore evaluates a relational structure already obtained from the text. Its later success is a further observation concerning that structure. The research history determines which relationships were available at each stage. The final probability experiment described here takes the completed catalogue as its input and evaluates one complete commitment without calendar-guided revision. The chronology of the earlier recognitions remains part of the provenance of the discovery. The preliminary discussion expressed the probability as a product of a calendar frequency, an interval-selection factor, a topology factor, a pre-specification factor and an allowance for multiple attempts. Its equation was: p_early=M_eff p_calendar p_intervals p_topology p_prespec. The assigned values were ten effective attempts, a calendar frequency of 0.0002, an interval factor between ten to the negative sixteenth and negative eighteenth powers, a topology factor between ten to the negative eighth and negative twelfth powers, and a pre-specification factor between ten to the negative fourth and negative eighth powers. Their multiplication produced approximately two multiplied by ten to the negative thirty-first through negative forty-first powers. The midpoint was two multiplied by ten to the negative thirty-sixth power, whose reciprocal is five multiplied by ten to the thirty-fifth power. These were heuristic estimates. The interval, topology and pre-specification factors had no enumerated sampling distributions, and their separate contributions could overlap. The finite reference experiments developed below give explicit counting rules in place of those assigned multipliers. An earlier discussion also reported a random-shift simulation with a probability of approximately 0.00036, or one in 2,800, for a frozen twelve-event schedule. That result concerned random placement of an existing geometry. It receives no additional multiplier in the joint calculation, which evaluates calendar placement directly. A further preliminary shuffle experiment counted the permutations of fifteen distinct labels. Fifteen factorial equals 1,307,674,368,000, giving a probability of approximately 7.6472 multiplied by ten to the negative thirteenth power for one specified order. Permuting a list tests order alone. The endpoint experiment includes the shared boundaries, overlap and nesting present in the Capstone. The completed numerical catalogue contains fifteen primitive measures: 2,556; 2,520; 2,300; 1,335; 1,290; 1,260; 777; 450; 430; 150; 40; 30; 10; seven; and three and a half days. The network includes common origins, common termini, overlapping measures and intervals contained within longer intervals. The internal three-and-a-half-day position lies within the final seven-day witness week. These relationships require a model that permits the intervals to share event boundaries and to occupy positions within one another. Every measure carries its specified counting convention. In the arithmetic verified for the displayed 2028-2035 dates, the 1,260 and 1,290 are inclusive counts and correspond to 1,259 and 1,289 elapsed days. The 777th-day, 150th-day and fortieth-day requirements correspond to 776, 149 and 39 elapsed days. The verified specification uses 1,335 elapsed days and 430 elapsed days for the measures bearing those labels. The three-and-a-half-day measure is represented exactly by seven half-day units. Recording the convention with the measure preserves the same rule in every calculation. The network also retains the supplementary 490-day connection, the 69 elapsed days from the witness resurrection to the final boundary, the seven-day warning, the additional thirty-day relation, the separate three-day assertion and the six elapsed days spanning the seven festival dates of Sukkot. The implementation checks fifteen primitive measures and six supplementary or bridging assertions, giving twenty-one numerical checks. These checks all hold for the supplied offsets. Each check remains active even when another relationship already entails it. The construction calculation uses sixteen ordered event roles. An event role identifies a position such as the opening, the beginning of the shorter measures, the breach, the rebuilding call or the final boundary. Each pair of roles in chronological order is an eligible forward connection. Sixteen roles therefore provide 120 forward pairs: Q_s=(16¦2)=120. The reference experiment assigns each of the fifteen labelled measures to a different forward pair. The first label has 120 choices, the second has 119, and the remaining labels proceed through the unused pairs. Shared beginnings and endings remain available because two distinct pairs can contain the same role. Nesting and overlap remain available because the roles retain their chronological order while the selected connections can span one another. The resulting count is a falling factorial: N_s=(120)_15=120⋅119⋯106. This gives 6,185,983,879,158,893,706,209,144,832,000 complete labelled assignments. The family includes raw assignments whose numerical or textual requirements fail. The declared success event is the supplied source-labelled binding. The probability law gives every complete assignment equal weight. With one specified source-labelled target, its outcome fraction is: p_s=1/(120)_15 ≈1.6165577207×10^(-31). Uniform weighting is the defining rule of this reference experiment. The event catalogue, the interval values, their units and their conventions are fixed inputs. The calculation measures the specificity of the required binding within the stated family. In the general case, a success class containing several qualifying bindings has probability equal to its qualifying count divided by the total count: P_R ("success" )=K/N. The subscript identifies the reference law. The single-target result uses one qualifying binding. A probability experiment encompassing every comparably complete chronology requires the corresponding qualifying count for that broader class. The number of documented discoverers and the number of qualifying mathematical constructions are distinct quantities. The same counting rule applies to any source supplying the same catalogue, eligible connections and success requirements. Incomplete recognition can also enter the counting family. A reading that recognizes k of the fifteen measures first selects those measures and then assigns them to distinct forward pairs. Summing over all recognized subset sizes gives: N_all=∑_(k=0)^15▒(15¦k) (120)_k=7,120,988,781,466,615,731,459,670,826,641. This family contains the empty reading, partial readings and complete readings. Uniform weighting over this entire family gives approximately 1.4043 multiplied by ten to the negative thirty-first power for one specified complete target. A second law recognizes each measure independently with probability one half and then assigns recognized measures uniformly. Under that law, the complete-target probability is approximately 4.9333 multiplied by ten to the negative thirty-sixth power. A third law draws a shared recognition propensity uniformly between zero and one, then recognizes measures independently conditional on that propensity. It gives a full-recognition probability of one sixteenth and a complete-target probability of approximately 1.01035 multiplied by ten to the negative thirty-second power. All three laws permit overlooked intervals and differing connections. Their different probabilities identify the contribution made by the weighting rule. More generally, independent recognition with probability q for each measure gives: P_R ("complete target" )=q^15/(120)_15 . The principal joint calculation uses the complete-recognition law. The omission variants remain separately defined sensitivity experiments. Their factors describe alternative generation rules and are applied within their own experiments. The calendar calculation begins after the entire interval geometry has been frozen. It uses the fixed Hebrew calendar, including its leap-month pattern, variable Cheshvan and Kislev lengths, and postponement rules. The opening is 10 Tevet. The range from AM 4000 through AM 8999 supplies exactly 5,000 annual candidate openings. One opening is selected uniformly in the reference experiment. The fifteen interval values and all twenty-one numerical checks remain unchanged throughout the scan. The ten date-node targets are Tevet 10, Tishri 1, Tammuz 9, Av 10, Sivan 30, Tishri 15, Adar I 15, Tishri 10, Tishri 17 and Kislev 27, in their specified positions across the seven-year structure. These predicates identify exact Hebrew dates. They include festival positions and dates carrying the stated scriptural event correspondences. The rebuilding target uses Sivan 30, and the final Hanukkah target uses Kislev 27. Their specified Hebrew-year relationships are preserved. The first-Adar identity requires precise naming. February 24, 2035 is 15 Adar I, Shushan Purim Katan. Ordinary Shushan Purim in that leap year is 15 Adar II, March 26, 2035. The nonzero calendar results in this calculation use the first-Adar target. The ordinary Shushan Purim predicate is retained as a separately evaluated calendar definition and produces no complete ten-date matches for this frozen web in the scanned range. Calendar identity is part of the specification itself. [1] The joint scan finds 71 annual openings satisfying all ten exact date predicates. Requiring the span to equal seven Gregorian calendar years leaves 22. Requiring December 28 at both bookends leaves three. The corresponding calendar stages are: 5000→71→22→3. The three bookended windows are December 28, 2028 to December 28, 2035; December 28, 2400 to December 28, 2407; and December 28, 2620 to December 28, 2627. The ten-date-plus-bookends probability under the uniform annual-opening law is therefore: p_h=3/5000=0.0006. Every calendar condition is also evaluated separately. Starting with 5,000 Tevet-10 openings, Tishri 1 leaves 3,158. Tammuz 9 leaves 380, and Av 10 retains those 380. Sivan 30 leaves 283, and Tishri 15 retains those 283. The first-Adar-15 requirement leaves 71. The Tishri-10, Tishri-17 and Kislev-27 requirements each retain those 71. The seven-Gregorian-year condition then leaves 22, and the December-28 bookends leave three. This sequence records the actual reduction supplied by each additional condition. The probability of the combined conditions follows the multiplication rule for conditional probabilities. Each new requirement is evaluated among the outcomes already satisfying the preceding requirements: P_R (H_1,…,H_10∣S)=P_R (H_1∣S) ∏_(i=2)^10▒P_R (H_i∣S,H_1,…,H_(i-1) ). Here the fixed skeleton is the conditioning information. A retained fraction of one records a requirement that every remaining opening already satisfies. The requirement remains enforced. The thirty elapsed days from Tammuz 9 to Av 10 illustrate this dependence. Once the frozen thirty-day link and the Tammuz-9 landing hold, the Av-10 landing follows. Likewise, the frozen seven-day link carries Tishri 10 to Tishri 17. The construction experiment accounts for the selection of those connections; the calendar experiment evaluates their consequences under the fixed calendar. The fuller stored specification also requires Iyyar 21, Tishri 21, Cheshvan 10, Cheshvan 17 and Cheshvan 20 at the additional date nodes. All fifteen stored Hebrew profiles yield 38 annual matches. Requiring seven Gregorian years leaves eleven, and the December-28 bookends leave one. Thus the 71, 22 and three hierarchy belongs to the ten-target calendar definition, while the 38, eleven and one hierarchy belongs to the fuller fifteen-profile definition. The specified predicates determine which hierarchy applies. The macro chronology supplies a second network with seventeen ordered event roles. The dated catalogue runs from the Fall at 3966 BCE through the Flood, the patriarchal events, the Exodus, the division of the land, the temple periods and the first-century events, ending at 2035 CE. Its nine counted clock records are the 1,656-year genealogy total, the 430-year sojourning measure, the 400-year affliction measure, the 450-year land-division measure, the forty-year wilderness period, the second 430-year measure, one 490-year measure, the 3,430-year Jubilee chain and the 6,000-year framework. The supplied dates for these seventeen roles are the Fall at 3966 BCE, the Flood at 2310 BCE, Babel at 2209 BCE, Abram's birth at 1958 BCE, the covenant at 1876 BCE, the Isaac event at 1846 BCE, the Exodus at 1446 BCE, entry into Canaan at 1406 BCE, land division at 1396 BCE, temple construction at 966 BCE, temple destruction at 586 BCE, the decree at 457 BCE, Jesus' birth at 4 BCE, the crucifixion and resurrection at 31 CE, Stephen at 34 CE, Cornelius at 35 CE, and the final boundary at 2035 CE. These dates constitute the chronology's input catalogue. The combinatorial count evaluates the placement of clock labels among these roles; it carries the catalogue as given. The two 430-year records retain distinct source identities and event connections. The sojourning clock connects the covenant and the Exodus, while the Ezekiel-related clock connects the division of the land and the temple construction in the supplied chronology (Exodus 12:40-41; Ezekiel 4:5-6). The affliction and land-division measures retain their stated connections to the Isaac event (Genesis 15:13; Acts 13:17-20). The wilderness and seventy-week measures retain their corresponding event boundaries (Numbers 14:33-34; Daniel 9:24-27). The Jubilee and millennial totals are the derived long measures used by the macro framework. The resulting checked spans are 3966 to 2310 BCE for 1,656 years; 1876 to 1446 BCE for 430 years; 1846 to 1446 BCE for 400 years; 1846 to 1396 BCE for 450 years; 1446 to 1406 BCE for forty years; 1396 to 966 BCE for the second 430 years; 457 BCE to 34 CE for 490 years; 1396 BCE to 2035 CE for 3,430 years; and 3966 BCE to 2035 CE for 6,000 years. The nine counted records agree with the supplied year labels. The chronology record contains a further entry linking 457 BCE to a 35 CE endpoint with the same 490-year measure. At equal-phase year-label resolution, 457 BCE to 34 CE gives 490 elapsed years, while 457 BCE to 35 CE gives 491. That endpoint entry retains its unresolved phase status. One Daniel 9 measure enters the probability count, and the second recorded endpoint receives no additional clock factor. The present macro calculation therefore uses the nine checked clock records. Seventeen ordered roles provide 136 forward event pairs. The same uniform allocation rule assigns the nine source-labelled clock records to distinct pairs: Q_m=(17¦2)=136, N_m=(136)_9. There are 12,145,555,288,464,998,400 such assignments. With one specified source-labelled macro target, the corresponding outcome fraction is: p_m=1/(136)_9 ≈8.2334646399×10^(-20). The specified source identities distinguish the two 430-year labels. An arithmetic-only acceptance rule that exchanges those labels while preserving all numerical differences admits both exchanged assignments. The one-target calculation retains the textual event identities as part of its success criterion. The absolute date catalogue is supplied to this experiment; the randomized object is the assignment of its clock records to the eligible event connections. The two long macro chains reach the same year. The Fall anchor of 3966 BCE and the 6,000 elapsed years reach 2035 CE. The land-division anchor of 1396 BCE and the 3,430 elapsed years also reach 2035 CE. The era calculation accounts for the absence of a year zero: 3966+2035-1=6000, 1396+2035-1=3430. The two equations impose one shared terminal-year condition once their anchors and lengths are supplied. Of the three December-bookended calendar windows, one ends in 2035. Conditional on those three windows, agreement with the macro endpoint therefore contributes: P_R ("2035 agreement" ∣"bookended snap" )=1/3. The combined calendar and endpoint weight is consequently the product of the three-in-5,000 bookended rate and the one-in-three endpoint selection: 3/5000⋅1/3=1/5000. At the remaining 2028 opening, the five additional stored Hebrew profiles also hold. Applying those profiles after the 2035 condition retains the same single opening. Applying them before the macro-year condition likewise leaves the same opening. The joint calculation records each required check while assigning one weight to their shared reduction. Both long clocks receive their structural binding factors within the nine-clock count, and the shared terminal year receives one calendar-phase factor. The complete reference experiment now consists of a micro binding, a macro binding and one annual calendar opening. Its joint law gives uniform weight to the stated binding families and to the 5,000 openings. This product law specifies how the modeled outcomes are generated. It supplies a conditional probability calculation for the exact target under those rules. The historical probabilities of source composition and of individual reading habits remain separate empirical questions. The three principal components can be written as the skeleton probability, the bookended holy-day alignment probability, and the macro probability conditional on agreement with that calendar outcome: p_s=1/(120)_15 , p_h=3/5000, p_(m∣h)=1/(3(136)_9 ). Their product gives the probability for one complete blind commitment to the joint target: p_one=1/(120)_15 ⋅3/5000⋅1/(3(136)_9 )=1/(5000(120)_15 (136)_9 ). Numerically, this is: p_one≈2.6619741664×10^(-54). The reciprocal is approximately one in 3.76 multiplied by ten to the fifty-third power. This value applies to the reference experiment with all nine macro bindings included as random allocations. It combines the structural selections with the complete required calendar phase. The term assigned to the calendar already contains the individual date conditions, the seven-year span, the December bookends and their agreement with the 2035 macro endpoint. Historical opportunity enters through a population allowance. A cumulative census of unique individuals with access to the relevant passages is unavailable in the research record. The calculation therefore adopts an intentionally broad allowance of 63 billion modeled individuals. The Population Reference Bureau's published estimates imply approximately 62.0 billion births between 1 CE and 2022, with approximately 300 million people alive at the beginning of that period. Their sum is about 62.3 billion. Rounding upward to 63 billion provides an overinclusive benchmark relative to that demographic estimate and includes people beyond the actual source-access population. [2] Each modeled individual receives one complete blind opportunity. Let the event for an individual be that the same person obtains the required micro binding, macro binding and calendar agreement. The probability that at least one individual succeeds is bounded by the sum of the individual success probabilities. This is the union bound: P_R ((⋃┴M)┬(j=1) E_j )≤∑_(j=1)^M▒P_R (E_j )≤Mp_one. This inequality accommodates dependence between people. Its application requires the stated upper bound on each person's modeled marginal probability. With the population allowance fixed at 63 billion, it yields: P_R ("at least one joint success" )≤(63×10^9 )(2.6619741664×10^(-54) )≈1.6770437248×10^(-43). The reciprocal of this upper bound is approximately one in 5.96 multiplied by ten to the forty-second power. The population multiplier appears once because the event concerns one person obtaining the joint result. The modeled opportunity is one full commitment per person. A law for repeated lifetime attempts would define a different per-person success probability. The same population allowance can be applied to the components separately. The complete micro binding alone has an upper bound of approximately 1.0184 multiplied by ten to the negative twentieth power. The micro binding together with one uniform opening satisfying the strict calendar phase has an upper bound of approximately 2.0369 multiplied by ten to the negative twenty-fourth power. The nine-clock macro target alone has an upper bound of approximately 5.1871 multiplied by ten to the negative ninth power. These are component results under their respective reference laws. The joint result uses the population allowance once for the same person's complete success. A further reference experiment searches the whole 5,000-year range after drawing the two bindings. Under the single-target success definition, the required calendar fit is then a property checked inside the target. The micro and macro target probability is approximately 1.3310 multiplied by ten to the negative fiftieth power, with a 63-billion-person upper bound of approximately 8.3852 multiplied by ten to the negative fortieth power. The additional one-in-5,000 factor belongs specifically to the experiment that draws one annual opening. Stating the opening rule distinguishes success somewhere in a search range from success at one randomly drawn opening. The macro model can also be conditioned on the two long-clock bindings as already supplied. Holding their complete source-to-final connections fixed leaves seven labels and 134 unused forward pairs. The remaining assignment count is 661,522,619,197,440. In that conditional form: p_(m∣two long bindings)=1/(134)_7 ≈1.5116641079×10^(-15). Retaining the same annual-opening experiment and population allowance gives a joint population upper bound of approximately 3.0790522788 multiplied by ten to the negative thirty-ninth power. The full nine-binding calculation and this conditioned calculation specify different amounts of information as already given. Their separate definitions account for their different numerical results. The computational record supports each arithmetic layer directly. Python verifies the fixed offsets, every stated numeric assertion and the exact assignment counts. JavaScript independently verifies the combinatorial fractions using integer arithmetic. The joint calendar verification converts 80,000 projected dates through the Intl and ICU Hebrew-calendar interface, compares all fifteen stored date predicates for each annual opening, and reproduces the matching-year lists. These checks establish reproducibility under the stated specification. [3] The probability law determines the meaning of the final number. The graph counts are exact within their defined families, the calendar counts are exhaustive within the 5,000 annual openings, and the combined fraction follows from the stipulated joint weighting. The interval values, event catalogues, chronology anchors and success criteria are the conditions of that calculation. The observed frequency of human interpretations requires a corresponding empirical distribution over their possible outputs. Historical novelty remains recorded as provenance, with its numerical contribution confined to any separately defined historical model. The discovery argument also includes the ancient origin of the interval texts, the subsequent development of the fixed calendar, the thematic correspondences and the absence of a documented earlier complete reconstruction in the supplied research record. The calculation retains these features as historical and textual evidence. The reference law assigns numerical weights to the defined bindings and calendar openings. A separate numerical factor for the historical sequence would require its own distribution, alongside its relationship to the factors already counted. A comparison between chance and intentional ancient coordination uses likelihoods under the competing explanations. If the observed evidence is denoted by E, intentional coordination by D and the specified chance explanation by C, the posterior odds obey: P(D∣E)/P(C∣E) =P(E∣D)/P(E∣C) ⋅P(D)/P(C) . The joint outcome fraction calculated in this chapter supplies a probability under the declared reference law. A numerical posterior comparison additionally requires predictive likelihoods for the causal explanations and their prior weights. Those inputs retain their own definitions and evidential requirements. The reported result remains the exact conditional calculation for the specified target: approximately 2.662 multiplied by ten to the negative fifty-fourth power for one blind commitment, and an upper bound of approximately 1.677 multiplied by ten to the negative forty-third power across the 63-billion-person allowance. The record therefore preserves the textual construction, the calendar agreement and the macro convergence as identifiable parts of the evidence. Each required relationship remains visible in the calculation. The stated sample spaces, conditional factors and shared endpoint give the numerical results their fixed mathematical meaning. Source notes [1] Hebcal, “Shushan Purim Katan 2035” and “Shushan Purim 2035.” The listed civil-day endings are February 24 and March 26 respectively. https://www.hebcal.com/holidays/shushan-purim-katan-2035 ; https://www.hebcal.com/holidays/shushan-purim-2035 . [2] Toshiko Kaneda and Carl Haub, “How Many People Have Ever Lived on Earth?” Population Reference Bureau, November 15, 2022. The population allowance uses the estimated births after 1 CE and the population alive at that date. https://www.prb.org/news/how-many-people-have-ever-lived-on-earth/ . [3] Companion calculation record, capstone_constraint_reductions.zip, archive version 10, November 3, 2026. Relevant records include INTERPRETATION_SPACE_REPORT.md, macro_endpoint_extension/MACRO_ENDPOINT_EXTENSION_REPORT.md, calendar_joint_snap/CALENDAR_JOINT_SNAP_REPORT.md, their input specifications, executable calculations and independent verification records. The discovery chronology draws on the May 23, 2013 research text reproduced in the research account.