Market Data in Each Run

Every FinZot run, meaning the detailed projection and every Monte Carlo trial, draws each year’s returns, distributions, and inflation from one built-in catalog of historical market data. What differs from run to run is which slice of that catalog a given year reads and which overrides are allowed to rewrite it. This page maps how the detailed projection and the trials are priced. It also states the fallback FinZot applies when an input is missing or a fund’s history is too short. What the detailed projection and the trials fundamentally are, and what the success rate measures, is covered in Projections & Monte Carlo. The full rule for which runs an override reaches is covered in asset return overrides.

Two independent settings

FinZot sets the detailed projection’s source and the way the trials are built independently. Changing one leaves the other’s results intact.

• The detailed projection is the single run reported in full, year by year. Its source is one of the three below.
• The trials are generated separately. The block bootstrap that builds them is a different setting entirely, and it never draws from the detailed projection’s source.

The detailed projection’s three sources

The detailed projection reads its year-by-year returns one of three ways.

Source	What the detailed projection reads each year
Expected returns, the default	Each held fund’s fixed long-run average return over its own full history, the same every year. If the fund has a return-rate override, that rate instead (returns and distributions, asset return overrides).
Historical replay	A named window of history spliced in starting at a trigger event, then looped or reverted afterward. You choose whether it repeats N times, repeats indefinitely, or reverts to your expected returns. It replays each year’s actual return, so a return-rate override does not apply.
Monte Carlo trial	One trial replayed in full detail. Plan year N replays the N-th sampled year, recentered on any return-rate override exactly as the trial itself was.

Uncovered years fall back to expected returns. Any plan year a historical replay or a replayed sequence does not itself cover uses the expected-returns path, meaning the same long-run average returns or the override in their place. This covers a replayed sequence shorter than the projection, the years before a replay’s trigger fires, and the years after its sequence finishes.

How the trials are generated

The trials come from a block bootstrap. It draws one sequence of historical years per trial, made of consecutive slices of real history stitched to the plan’s length. So every trial is, in effect, a replay over sampled years. A fixed random seed makes the trial set reproducible. The trial count, block length, and how the survivor fraction becomes the success rate are covered in Projections & Monte Carlo.

Which runs an override reaches

A scenario can set a return-rate override or a tax-mix override on a fund. Which runs each one reaches, including how a return-rate override is recentered on a trial’s sampled history, is covered in asset return overrides. In short, a return-rate override reaches the expected-returns path and every trial, recentered, including a single trial replayed in full detail. It does not reach a historical replay. A tax-mix override reaches every run. What this page adds is how those inputs meet the fallbacks below when an override is absent or a fund’s history is too short.

Fallback A: a missing override means default, not zero

An override’s absence is never treated as zero. It means “use the fund’s own data.”

• No return-rate override. FinZot uses the fund’s base return and yield, from its own history.
• No tax-mix override. FinZot uses the fund’s base tax split, the one embedded in its historical series.

Fallback B: pre-history for years before a fund’s real record

Any year priced off a specific past year, whether a historical replay year or a trial year, can ask for a year before a fund’s own market record begins. A short-history catalog fund carries its own reconstructed pre-history back to the catalog’s start in 1871. Those years are the fund’s own data, correlated with a long-history proxy but not identical to it, so the fund behaves like a full-history fund rather than a copy of the proxy. Every catalog fund reaches that same start year. So trials can sample the full window no matter which funds a plan holds, and no held fund is ever missing a year. A custom asset inherits its parent’s full series and reaches back just as far. The expected-returns path never reads an individual past year at all. It reads each fund’s fixed long-run average, or its override.

Reconstructed years carry the fund’s overrides too. A tax-mix override re-slices the fund’s distributions in its reconstructed pre-history just as in its observed years. Even where a year’s yield is modeled rather than observed, your split is applied to it, so the tax type stays consistent across every year the plan draws on. A return-rate override likewise recenters a trial’s sampled returns whether the sampled year is observed or reconstructed.

How a single year’s return and distribution are composed, before any of these overrides apply, is described in returns and distributions. How a custom asset inherits its parent’s series is in custom assets.