01 · Frame
A narrow question, held every day
Bits.Bid does not attempt to predict every market move. It does not scan thousands of tokens. It does not execute trades.
Instead, it maintains a disciplined daily Run Book, evaluates a defined body of evidence, and turns that evidence into a relative view between two assets.
One asset may get the bid.
The allocation may hold.
Or there is no clean edge.
Those three outcomes are the product. “No clean edge” is a credibility feature, not a failure state.
02 · Run Book
What the Run Book does
The Run Book is the research engine underneath the daily Bid. Its purpose is to make the decision process repeatable rather than emotional.
The underlying system was originally developed as a deterministic AVAX research framework. It collects market and network information, derives defined metrics, evaluates explicit states and gates, stores each observation, and preserves the history needed to replay previous runs. Its architecture deliberately keeps generative AI outside the calculation path so the same inputs can reproduce the same result.
Bits.Bid expands that discipline from:
“Does the evidence support adding AVAX?”
toward:
“Between AVAX and BTC.b, which asset currently deserves more of the allocation?”
The comparison becomes the product. The Run Book remains the structural evidence layer — it does not answer the pair question on its own.
03 · Evidence
The evidence is evaluated in layers
Bits.Bid does not rely on one indicator. The methodology considers several classes of evidence because they answer different questions.
Structural evidence
Structural evidence asks whether the underlying economic case for an asset is strengthening or weakening.
Transparency is only worth something if you can tell what actually moved today's number. So the inputs are split: what the current model scores, and what the research programme is still working on. Only the first column can change a published bid.
What scores today
Reads directly into the published score.
- Capture state — is value accruing to AVAX itself
- Supply state — supply economics and dilution
- Economy state — network economic health
- Momentum — AVAX against BTC over 7 / 30 / 90 days
- Canonical regime — the risk environment around the pair
- Realized volatility — trims the tilt into an allocation
The structural tilt is exactly capture + supply + economy + catalysts, clamped to the published range. Every one of these appears in the daily record.
Research frontier
Studied, not scored. None of these can move today's bid.
- Network activity as a distinct series
- Fee and burn dynamics
- Ecosystem development
- Additional structural evidence
- Additional asset matchups
- Challenger signals
These inform the Run Book's written analysis and future methodology versions. They are not wired into the score.
One caveat, stated plainly. A catalyst term exists in the scoring formula and is bounded at ±1 point, but no data source currently feeds it, so it contributes 0 to every published bid. You can verify this in any daily record: the structural reason line ends with catalysts 0. It is listed above as scoring rather than frontier because the code path is live — but it is inert until a source is wired in.
A key principle: activity alone is not treated as value capture. A network can become busier without that activity improving the economics of its native token. That distinction is central to the original AVAX Run Book methodology — and it is why network activity sits in the frontier column rather than the scoring one.
Relative market evidence
Bits.Bid is not evaluating AVAX and Bitcoin in isolation. It is evaluating them against each other.
Relative market evidence examines whether AVAX is strengthening or weakening relative to Bitcoin across defined periods. The product question is not simply “Is AVAX going up?” It is:
Does AVAX currently have a stronger relative case than BTC.b?
Published cryptocurrency research has found meaningful evidence of momentum in crypto returns. Liu and Tsyvinski’s Risks and Returns of Cryptocurrency documents strong time-series momentum effects, while Liu, Tsyvinski and Wu’s Common Risk Factors in Cryptocurrency identifies momentum as one of the significant factors associated with the cross-section of cryptocurrency returns.
Bits.Bid uses this research as methodological support for considering trend and relative strength. It does not imply that past momentum guarantees future performance.
Market regime
The same signal can have different meaning under different market conditions. Bits.Bid therefore considers the broader environment around the two assets rather than assuming all periods are equivalent.
A risk-off environment, for example, may affect how much evidence is required before increasing exposure to the more volatile side of the pair. This is a risk-control input, not a prediction that any particular market regime must continue.
Risk and volatility
Return is only one side of allocation. Risk matters too.
Bits.Bid can consider differences in volatility, drawdown and related risk characteristics when translating evidence into a portfolio balance. The purpose is not automatically to choose the least volatile asset. It is to avoid treating two assets with materially different risk profiles as though equal evidence should always produce equal position changes.
Academic work on volatility-managed portfolios provides precedent for changing risk exposure as volatility changes rather than maintaining constant risk through every market environment. Moreira and Muir, Volatility-Managed Portfolios, is an important example.
04 · The Bid
From evidence to the Bid
The methodology separates two questions — and, as of the current model, two product fields.
Who has the edge?
The research engine compares the evidence supporting AVAX with the evidence supporting BTC.b. That produces the daily relative preference: AVAX gets the bid, BTC.b gets the bid, or no clean edge.
What should the book do?
A separate field — STANCE — answers how the portfolio should move, including toward or away from cash. The Bid never sizes cash. Cash is an overlay, never a bid candidate.
The Bid is a relative research conclusion. It does not mean an asset is guaranteed to rise. It does not mean the other asset is expected to collapse. It means that, under the current methodology and current evidence, one side has earned the stronger relative case.
Is the edge strong enough to change the balance?
A new Bid does not necessarily require an immediate portfolio change. Bits.Bid distinguishes between signal and action.
Before changing the model balance, the methodology can consider:
- strength of the relative edge
- existing allocation
- market risk
- persistence of the signal
- trading friction
- whether the change is large enough to justify a rotation
This is intentional. Repeatedly rotating a portfolio because of small changes in noisy signals can destroy value through turnover and transaction costs. Research into portfolio rebalancing with transaction costs supports explicitly considering those costs rather than treating trading as frictionless.
05 · Portfolio
How Bits.Bid reaches a portfolio balance
Bits.Bid does not pretend that portfolio construction can be known to two decimal places. The methodology translates the strength and quality of the evidence into a target allocation state.
A stronger AVAX case can justify an AVAX overweight. A stronger BTC.b case can justify a BTC.b overweight. Mixed evidence can support a more balanced state. An edge that has not changed enough can leave the existing allocation untouched.
Cash exposure is controlled separately by the overlay: absolute-risk signals (regime, conviction, peg) can raise or lower total risk in the book without rewriting who won the relative Bid. A rotation leaves total exposure unchanged; a move to or from cash changes how much is at risk.
This follows the broader principle introduced by Harry Markowitz’s foundational Portfolio Selection: investment decisions should be considered at the portfolio level, with attention to both expected outcomes and risk rather than evaluating each asset independently.
Bits.Bid is much narrower than modern portfolio theory — it deliberately compares only two assets in its initial form — but the portfolio-level way of framing the decision is well established.
06 · Reproducibility
Why the methodology is deterministic
The calculation path is designed to be reproducible. The daily result should come from defined data, defined transformations and defined decision rules.
The methodology should be able to explain why the Bid happened.
This makes historical evaluation possible. A previous day’s result can be preserved with the inputs that produced it and compared with what happened afterward.
The Run Book already stores append-only observations, supports historical replay and calculates counterfactual outcomes against AVAX, BTC and cash. Bits.Bid builds on that discipline. There is no model provider, no prompt and no temperature in the scoring path — by design.
07 · Boundaries
What Bits.Bid is not
- Not an exchange
- Not a wallet custodian
- Not a trade placer
- Not a controller of user funds
- Not a promise of returns
- Not a claim of certainty
And it is not designed around an AI model making unexplained investment decisions.
Bits.Bid is intended to make a defined research methodology visible, repeatable and measurable.
08 · Research
Research, not certainty
Bits.Bid is research-oriented and draws on concepts studied in academic finance and cryptocurrency research. Bits.Bid itself is not a peer-reviewed academic study.
Published evidence supporting momentum, portfolio construction, volatility management or cost-aware rebalancing does not prove that the Bits.Bid methodology will outperform AVAX, Bitcoin, a 50/50 portfolio or any other benchmark.
That question must be measured.
The product therefore treats historical results, counterfactuals and future observations as part of the research process rather than using academic citations as a claim of future performance.
09 · Transparency
Methodology transparency and proprietary IP
We believe users should understand what kinds of evidence matter and why they matter.
We do not believe public methodology disclosure requires publishing every implementation detail.
Public methodology may describe
The research framework, signal categories, portfolio logic, risk principles, data provenance, evaluation process and reasons behind a Bid.
It does not disclose
Exact weights, thresholds, transformations, scoring functions, calibration techniques, internal decision boundaries, source-selection logic or other implementation details that constitute Bits.Bid intellectual property.
Transparency should make the result understandable without making the proprietary engine fully reproducible by a competitor.
Important notice
Model research. Not advice.
Bits.Bid provides model-based research and educational information.
Crypto assets are volatile and can lose substantial value. A Bid is not a promise of performance, a guarantee, or an instruction to transact.
Users remain responsible for their own financial decisions, tax considerations, custody choices and risk.
Methodology descriptions are provided to explain how Bits.Bid approaches research. They should not be interpreted as proof that any methodology will produce profitable results.