Rival reconstructions

Every timeline at once. Decide for yourself.

The usual way to draw a disputed life is one track with fuzzy edges — birth somewhere in here, death somewhere in there. It is the wrong picture, and this site has already been caught making the same mistake in another form: averaging across the three Nisan rules produced a number describing no scholar's position, because the rules cannot all be true at once. A blurred band does that in graphics. It draws a smear nobody holds and calls it the state of the question.

So here is every serious reconstruction as its own track, each one internally coherent, each carried through from birth to death — with what it requires you to accept, what it costs you, and how it scores. They are ordered by how much of the textual evidence space lands on them. That ordering is not a probability and the strongest track is not the answer. It is a map of where the defensible readings cluster, and the trade each position makes is written out so you can weigh it yourself.

What is drawn and what is not. Everything here is at year resolution, because that is the resolution the evidence has. The gospels do not support a within-year itinerary; they disagree with each other on the order of events; and any month-level position for anything other than the death would be invention. The death dates alone are exact, because those are computed from the moon. Birth windows are the softest thing on the page — see the caveat below the chart.

The tracks

birth window (Luke 3:23) the part of it that clears Herod ministry Passover fixed point

The birth bars are the softest thing on this page, and one of them is circular. There is no independent birth datum here at all. Each window is derived from Luke 3:23 — "about thirty years of age" — read as thirty give or take 3. That sentence is in the same paragraph as Luke 3:1, by the same author. So a track whose birth window supports its ministry date has not corroborated anything: it is one source counted twice. Widen "about" and the bars swallow every proposal; narrow it and you have invented a precision the Greek does not carry. Treat them as a consequence of each position, never as evidence for it.

What the chart shows that the arguments usually don't

Two results fall out of drawing them together, and neither is visible when the positions are argued one at a time.

Matthew gets harder the later the ministry starts

Matthew 2 requires the birth before Herod died in 4 BC. Every track's birth window is a range, and only part of each clears that line — the gold portion of each bar. How much survives falls away steadily:

Track	Ministry begins	Birth window clearing Herod
AD 27	AD 25	83%
AD 30	AD 27	50%
AD 29	AD 28	33%
AD 33	AD 29	17%
AD 34	AD 30	0%

The gradient tracks the ministry, not the crucifixion — which is worth being exact about, because it is tempting to say "the later the death, the worse for Matthew" and that is not what the numbers do. AD 29 dies a year before AD 30 and fares worse, because it starts preaching a year later. What squeezes Matthew is the date Luke 3:23 is anchored to, and the squeeze runs all the way to AD 34, where nothing is left and the two gospels are simply incompatible.

Nothing reaches the census

Luke 2:2 dates the birth by the census of Quirinius, in AD 6. Not one track comes within 3 years of it — and they cannot, because the same gospels put the ministry where they do. The gap runs from 3 to 8 years across every reconstruction on this page.

This is drawn unreconciled on purpose. A timeline that split the difference between Herod and the census would have resolved by illustration what the field has not resolved by argument. The honest drawing is two markers that do not meet.

What each position costs

Ordered by share of the textual space. Read the right-hand column of each card before the left one.

AD 30, John's chronology

Friday 7/4 AD 30 · 14 NisanThe majority position in current scholarship.

Textual space11.4%Its visibility model vs P&D91.97%Age at death30–36Birth window7 BC – 1 BCMinistry2 years

What it requires

Tiberius reckoned so that the fifteenth year falls early — the co-regency, or a short ministry after a sole-reign start
The "fourteen years" of Galatians 2:1 running from the first Jerusalem visit, so that it adds to the three of 1:18
John's chronology: the death on 14 Nisan, as the lambs were killed

What it costs

The Sejanus argument, which would exclude this year outright, must be judged unpersuasive
The AD 33 eclipse, if it is Peter's "moon turned to blood", is not available to it
A three-Passover ministry starting in AD 27 sits awkwardly with a sole-reign reading of Luke 3:1

AD 29, the early-tradition reading

computes to Monday — not a Fridayastronomically refutedThe dominant date in several early Christian writers.

Textual space9.0%Its visibility model vs P&D91.97%Age at death28–34Birth window6 BC – AD 1Ministryunder a year

What it requires

The fifteenth year of Tiberius read straight off the sole reign, with a ministry of about a year
The synoptics taken at their surface, where a single Passover is described
Weight given to the patristic testimony as an independent line rather than as inference from the same gospels

What it costs

A one-year ministry is hard to reconcile with John, which names three Passovers explicitly
The early writers may simply be computing from Luke 3:1 themselves, in which case this is not independent testimony at all — it is the same argument with a longer beard
It fares poorly astronomically: 14 Nisan is not straightforwardly a Friday in AD 29 under the better-evidenced visibility models

AD 33, John's chronology

Friday 3/4 AD 33 · 14 NisanHoehner; Humphreys & Waddington.

Textual space3.2%Its visibility model vs P&D91.97%Age at death31–37Birth window5 BC – AD 2Ministry3 years

What it requires

Tiberius reckoned from the sole reign, putting the fifteenth year at AD 28/29
A ministry of about three and a half years beginning in autumn AD 29, requiring a fourth Passover that John does not name, inferred from the two-year gap between John 2:13 and 6:4
The "fourteen years" of Galatians 2:1 running from the conversion, so that it absorbs the three of 1:18

What it costs

The fourth Passover is an inference from a silence: John names three (2:13, 6:4, 11:55) and no more
It leaves a long gap between the death and Paul's conversion if the Gallio chain is read consecutively
Its two strongest supports — the eclipse and Sejanus — are both disputed inferences rather than data

AD 27, the sweep artifact

Friday 11/4 AD 27 · 14 NisanThis site, until the measurement caught it.

Textual space1.9%Its visibility model vs P&D89.17%Age at death29–35Birth window9 BC – 3 BCMinistry2 years

What it requires

The co-regency reckoning of Tiberius at its earliest
A crescent criterion that begins the month early

What it costs

It survives almost none of the textual space
It was this site's own headline result, produced because the least accurate visibility criterion occupied half the sweep and unanimously favoured it
It is kept on this page as a worked example of what a badly-weighted model space looks like from the inside

AD 34, the synoptic chronology

Friday 23/4 AD 34 · 15 NisanNobody argues for it. It is here because the calendar permits it.

Textual space0.5%Its visibility model vs P&D94.99%Age at death31–37Birth window4 BC – AD 3Ministry3 years

What it requires

The synoptic chronology: the death on 15 Nisan, after the Passover meal
A Nisan rule that pushes the month late
A late reading of the fifteenth year of Tiberius

What it costs

No published reconstruction defends it, which is itself evidence of a kind
It is admitted by well under one per cent of the textual space
It sits at the very end of Pilate's prefecture, leaving almost no room for the recall Josephus describes

Who actually holds what

The positions above are not anonymous. Below is who argues for each, what they argue from, and — the part usually left out — where we checked it. Every entry carries the URL consulted and the date, so you can check our checking.

These are marked secondary, a status that did not exist on this site until this section did. Elsewhere a citation is either verified (someone opened the edition) or unverified (written from memory). That binary hid a real third state: checked against an accessible source that reports the claim. It is a genuine improvement on memory and it is not the same as reading the book — a secondary source can misreport, and page numbers taken at second hand are the classic way an error travels. All 7 entries here are secondary. None is verified.

And a specific risk worth naming. Every entry states a real person's scholarly position. Getting one wrong is not like mis-citing a verse — it misrepresents an individual. So each is stated narrowly, in the terms the consulted source used, with a quotation where one was available. Where a scholar's view on one of this site's parameters is not attested in what was consulted, it is recorded as unattested rather than inferred. Reasoning from someone's crucifixion date to their reading of Galatians would be exactly the tidy fabrication this project exists to avoid.

AD 30

2 positions

John P. Meier

A Marginal Jew: Rethinking the Historical Jesus

Reconstructs from the fifteenth year of Tiberius, the start of John the Baptist's ministry, the length of the ministry and the years in which 14 Nisan fell on a Friday. Has the farewell meal on Thursday evening 6 April, the onset of 14 Nisan by Jewish reckoning, and Jesus dead by the evening of Friday 7 April.

Unattested in what we consulted: tiberiusEpoch, galatiansReading, passovers.

Position and reasoning summarised — source, consulted 2026-07-28.

Rainer Riesner

Paul's Early Period: Chronology, Mission Strategy, Theology (Eerdmans, 1998)

Reaches AD 30 by working back from the Gallio inscription through Acts and Galatians, and reports it as the majority view.

“the fourteenth of Nisan (7 April) of the year A.D. 30 is, apparently in the opinion of the majority of contemporary scholars as well, far and away the most likely date of the crucifixion of Jesus”

Unattested in what we consulted: galatiansReading.

The quoted sentence, reported — source, consulted 2026-07-28.

AD 33

2 positions

Harold W. Hoehner

Chronological Aspects of the Life of Christ (Zondervan, 1977)

The fullest case for AD 33. Builds from extrabiblical Roman material joined to the gospel data, especially Luke, and argues the AD 30 view is deficient. Requires the longer ministry.

Unattested in what we consulted: galatiansReading.

Date and method — source, consulted 2026-07-28.

C. J. Humphreys & W. G. Waddington

"Dating the Crucifixion", Nature 306 (1983), 743–746

Reconstructs the first-century Jewish calendar astronomically and identifies a partial lunar eclipse rising over Jerusalem that evening, calculated as visible between about 6:20 and 6:50 pm with the moon reddened, as the "moon turned to blood" of Acts 2:20.

Unattested in what we consulted: tiberiusEpoch, galatiansReading, passovers.

Journal, volume, pages and the eclipse timing — source, consulted 2026-07-28.

No single year

3 positions

E. P. Sanders

The Historical Figure of Jesus (Penguin, 1993)

Treats AD 30 as a useful approximation while stating plainly that a precise date cannot be had.

“In all probability Jesus was born in 5 or 4 BCE and died between 29 and 31 CE (though many scholars prefer 33)”

Unattested in what we consulted: tiberiusEpoch, galatiansReading.

The quoted sentence — source, consulted 2026-07-28.

Helen K. Bond — dissents

"Dating the Death of Jesus: Memory and the Religious Imagination", New Testament Studies 59.4 (2013), 461–475

Rejects the premise of precision. Argues the received dates rest on a chain of inferences the sources cannot carry, finds internal inconsistencies in Mark suggesting a death before Passover rather than on it, and holds that neither the Johannine nor the synoptic dating is secure enough to compute from.

“it is no longer a question of looking for 14th or 15th Nisan, because Jesus' death could have happened at any date round about then”

Author, journal, volume, pages, argument and the quoted sentence — source, consulted 2026-07-28.

David Instone-Brewer

"Jesus of Nazareth's Trial in the Uncensored Talmud", Tyndale Bulletin 62.2 (2011), 269–294

Not a dating argument. Establishes that the Munich manuscript of b. Sanhedrin 43a preserves material censored out of the printed editions, and that stratifying the tradition leaves as its oldest layer: "On the Eve of Passover they hung Jesus of Nazareth for sorcery and leading Israel astray." Later material, he argues, was added to resolve three difficulties — a trial on a festival, an unbiblical execution method, and the charge of sorcery.

“On the Eve of Passover they hung Jesus of Nazareth for sorcery and leading Israel astray”

Unattested in what we consulted: deathYear.

Journal, volume, pages, the reconstructed oldest layer, and the censorship claim — source, consulted 2026-07-28.

Read Helen Bond before you trust anything else on this site. She is the one entry that does not argue for a different year — she argues the question is not answerable at the resolution everyone else is working at. If she is right, the problem is not that this site picks the wrong candidates; it is that the whole exercise is mis-specified, because the gospels are not reliable enough at the level of the day to compute from at all. She finds internal inconsistencies in Mark pointing to a death before Passover rather than on it, and concludes the death falls somewhere between AD 29 and 34 around Passover, full stop. Her sentence — “it is no longer a question of looking for 14th or 15th Nisan, because Jesus' death could have happened at any date round about then” — is a direct denial of the premise of the sweep. It belongs here for the same reason the eclipse objections belong on their page: leaving out the strongest argument against what you are doing is advocacy, not teaching.

The earliest date, put through the engine

The AD 29 track above is astronomically refuted, and now it can be said exactly why rather than merely reported. Tertullian gives the earliest surviving explicit date for the crucifixion, and unusually he gives enough detail to test.

Tertullian, Adversus Judaeos 8.18

“sub Tiberio Caesare, consulibus Rubellio Gemino et Rufio Gemino mense Martio temporibus paschae, die octavo Kalendarum Aprilium, die primo azymorum quo agnum occiderunt ad vesperam”

Under Tiberius, in the consulship of Rubellius Geminus and Fufius Geminus — AD 29 — in March, at the time of the Passover, on the eighth day before the Kalends of April, which is 25 March, and on the first day of unleavened bread, when the lamb was killed at evening.

Two checks, and they disagree with each other:

The weekday holds

25 March AD 29 was in fact a Friday. Whatever else is going on, the date is not careless about the day of the week — which is more than can be said for a good deal of ancient chronography.

The Passover cannot

It cannot be Passover. The equinox falls around 22 March, so under every Nisan rule this project implements the month begins weeks later: our engine puts 14 Nisan AD 29 on 18/4, a Monday, 25 days after Tertullian's date.

So the earliest patristic date is internally consistent about the weekday and astronomically impossible as a Passover. That is a real result and it points somewhere: its author was computing from something other than the moon. 25 March carries independent symbolic weight in early Christian chronography, and a date arrived at symbolically will land on a plausible weekday while missing the lunar calendar entirely — which is precisely the pattern here. It is also the concrete reason the AD 29 reconstruction fails, rather than merely being unfashionable.

The one number nobody publishes with a range

Every reconstruction implies an age at death, and almost no treatment states it with error bars. Across all five tracks it runs from 28 to 37 — and the width of that range is not measurement error. It is the width of the disagreement about Luke 3:23's "about", stacked on top of the disagreement about the year. The familiar "thirty-three years old" is one point inside it, and it is the point you get by making a particular set of choices this page has now made visible.

Read this before treating the ordering as an answer. One: the tracks are ranked by share of the textual configuration space, which measures where defensible readings cluster and nothing else. It is not a probability, and the top track is not "the correct one". Two: "what it costs" is hand-written. The scores are computed, the trade-offs are a judgement, and pretending to compute them would have been worse than saying so. Three: the birth windows are derived from a sentence by the same author as the ministry date, so they corroborate nothing. Four: all 21 citations behind this page are marked unverified — the arithmetic is tested, the editions have not been opened. Five: a position being astronomically refuted means the calendar excludes it under the best-evidenced models this site implements, which is a claim about models, not a proof about history.

Where the textual shares come from → · Where the astronomy comes from → · Sources →