We do not predict markets.
We balance exposure against them.
G-BOT is a private, decades-in-the-making algorithmic trading initiative
for professional investors and institutions. There are no management
fees, no licences, no signals to subscribe to. You evaluate the
methodology yourself, on a paper-trading account, with live tick data,
for as long as you like.
The system is not based on forecasting or directional inference, but on
continuous control of a path-dependent inventory process driven by
Historical Trading Information (HTI), where every fill, hedge, and roll
becomes state.
"You just want to make money and balance exposure. Everything else
that is sold to retail traders — signals, technical analysis, forecasts
— is, statistically, noise."
The edge is not in the chart. It is Information (HTI).
G-BOT is engineered on the Universal Statistical Edge (USE)
framework—the culmination of decades of research into the mechanics of
compounded returns. We reject conventional trading heuristics: forecasting,
standard technical analysis, signal-based entry, and price pattern recognition
are explicitly excluded from our execution logic.
The system operates as a path-dependent state machine where every execution,
hedge, and rollover is recorded as Historical Trading Information (HTI),
continuously conditioning future inventory behaviour and risk constraints.
USE Principle
Markets do not have to mean-revert. Your positions must.
Our edge is decoupled from underlying asset volatility. Instead, it relies on
proprietary inventory cycles transitioning through harvest and reset states,
anchored by Historical Trading Information (HTI) and induced price recurrence.
Execution is not driven by directional prediction but by continuous
management of aggregate exposure across the portfolio, where all
risk contributions (Delta, Gamma, Vega, Theta and cross-effects) are
implicitly embedded in the portfolio-level expansion dynamics.
Hedging is therefore not a separate decision layer, but a mechanical
consequence of maintaining bounded exposure under real-time margin and
risk constraints.
Historical Trading Information is necessary for an edge.
Every fill, every stranded order, every rollover is preserved as structured memory.
The algorithm does not "predict" the next tick. It reconstructs context:
where the inventory originated, under which risk state it was opened,
and under what conditions it can be efficiently closed or rebalanced.
HTI is therefore not historical logging, but a dynamic state-conditioning layer
that defines future execution and hedging paths.
Bull and bear are positions in a balance, not a forecast.
G-BOT manages bilateral inventory whose total exposure is bounded
by hard caps and rebalanced continuously. Asymmetry is corrected
mechanically, not narratively.
What matters is not the sign of individual Greeks, but the net
aggregated portfolio response to infinitesimal changes in spot,
volatility, and time—i.e. the effective exposure of the system
under real-time market conditions.
Eight principles you will recognise from the first session.
We do not try to be clever about the future. We try to be disciplined
about what we already hold. The same set of ideas reappears whether you
load one instrument or fifty.
The actual source of edge is not prediction, but the controlled evolution
of a path-dependent inventory system (HTI) under binding risk constraints,
where hedging flows themselves become the mechanism through which decay,
imbalance, and volatility are structurally harvested.
01
No signals. No forecasts. No technical analysis.
Every chart pattern, every indicator, every momentum signal is
rejected as a primary trigger. The Kaufman SDX tracker is used only
as a veto gate, not as a buy or sell signal.
02
Inventory as a controllable price loop.
Each instrument is a Layer. Each Layer contains Players — units
of inventory with a side, an average entry, and a lineage. The bot
opens, hedges, repairs and rolls Players to harvest decay and reset
conditions.
Hedging is not a single action but a continuous control process
composed of three interacting mechanisms:
delta-hedging via liquid futures/spot proxies,
gamma-driven rebalancing (gamma scalping as a flow consequence, not a signal),
and implicit vega control through volatility exposure normalization.
03
Rollover is information transfer, not exit.
When a Player rolls, its identity moves to the new leg. The
Historical Trading Information follows. The price loop does not
depend on the underlying actually mean-reverting; it depends on your
own controllable recurrence under HTI state continuity.
04
Hard caps are non-bypassable.
Greek budgets, exposure caps and margin fractions block new risk
unconditionally. Closes, forced flatten and risk-reducing hedges are
always allowed. "Close-before-open" is sacred.
The system edge emerges precisely from operating under these constraints:
constrained hedging forces structural rebalancing asymmetry, which is
where realised PnL originates.
05
Filters before fills.
Crossed quotes, stale quotes, out-of-line bid or ask, thin
top-of-book sizes — all of them are refused. The filter set is the
single most rewritten part of the codebase. The defaults are
paranoid and correct.
06
The operator is in charge.
Every Layer has a Manual / Auto switch. Every order goes through
an order book the operator can read. There is no "flatten the whole
folio" button by design — if you need to flatten, you flatten Layer
by Layer while watching each fill.
07
Capital is part of the strategy.
The system is engineered to scale — single accounts up to roughly
$500 million, multi-account architectures beyond. Sub-capitalised
deployment will simply not work — this is the only honest answer.
08
Volatility is opportunity, not threat.
The biggest absolute profits come from movement, not calm.
Quiet markets produce few opens by design. Storms produce the most
information — if your filters and caps are respected.
Volatility is not “traded”, it is internalised through hedging flow
imbalance correction across the portfolio.
// 03 · Operating model
How a session actually works.
G-BOT runs on your hardware. It talks to Interactive Brokers via the
IB Gateway over a port you choose. Your capital never leaves your IBKR
account. We have no read access, no write access, no view of your
positions.
The broker connection is only the execution transport layer. The actual
system behaviour is determined by a higher-level control architecture
composed of three interacting subsystems: inventory state (HTI),
risk/margin constraints, and continuous hedging feedback loops.
A single secure link.
The trading executable communicates directly with the IB Gateway
over an encrypted, user-defined port. Tick data, Greeks, fills,
rejections and margin events flow on that link and nowhere else.
These inputs are not used as standalone signals, but as real-time
state variables feeding a portfolio-level control system that
continuously recomputes aggregate exposure and hedging pressure.
This is not "privacy theatre." It is an architectural choice:
we do not collect what we do not need.
What we don't know we cannot betray. What we don't store we cannot
disclose.
Anonymity is permitted.
During the private evaluation phase you may use a pseudonym. You
are under no obligation to reveal your real identity to us.
You evaluate the methodology on your own IBKR paper-trading
account with real-time data. You watch every fill, every Greek,
every cap event. There is no obligation to ever move to real money,
and no time limit on this stage.
Step 2 · Real-money trading
Same software. Same console. Your capital, your control.
Once you have understood the methodology and verified forward
performance with your own eyes, you switch the same executable to a
real-money account. There is nothing to license, nothing to install
on our side, and no fees to pay.
Step 3 · Scale
Multi-account architecture from one console.
Single instances cap practically around half a billion USD for
clean risk segmentation. Beyond that, multiple instances target
separate accounts from the same operator workstation — designed
for family offices and institutional advisors.
Scaling does not increase “strategy complexity”, it increases the
dimensionality of the hedging and margin-constrained rebalancing
system.
Caveat · Capital floor
Minimum $700,000. Optimal from $3 million.
The instruments G-BOT trades require adequate capital to keep
Greek budgets, exposure caps and margin headroom in healthy
territory. Under the floor the system simply will not generate
meaningful work.
The key constraint is not entry execution, but the ability to
sustain continuous hedging under margin and exposure limits
without destabilising the inventory state machine.
// 04 · Transparency
Don't take our word for it. Watch it run.
We do not publish back-tests because, in a dynamically hedged system, they are not a valid representation of execution reality.
Any simulation that replays historical prices without reconstructing real-time margin constraints, Greeks evolution, and inventory path-dependence
produces a structurally incomplete model of risk and return.
Why backtests fail in this system
Static price replay cannot reconstruct a dynamic hedging engine.
Standard backtesting assumes that price paths are sufficient to reconstruct PnL. This breaks under any system where:
(i) exposure is continuously rebalanced,
(ii) options Greeks evolve intraday and interact with inventory,
(iii) margin requirements change in real time,
and (iv) execution decisions depend on the full Historical Trading Information (HTI) state.
In practice, a historical price series does not contain the information required to reconstruct:
real-time gamma exposure, margin compression effects, hedging feedback loops, or forced liquidation thresholds.
As a result, backtests tend to produce smooth equity curves that are mathematically valid under simplified assumptions,
but economically non-actionable in live trading.
Path dependence & USE principle
The edge is not in the sequence of prices, but in the sequence of states.
The Universal Statistical Edge (USE) framework is explicitly path-dependent:
every execution decision is conditioned on the system’s own historical trading information (HTI),
including prior hedges, residual inventory, realized decay, and remaining Greek exposure.
This implies that two identical price paths can produce entirely different outcomes depending on internal state.
Backtests collapse this state dimension and therefore remove the actual mechanism that generates or destroys edge.
Gamma hedging reality
Greeks are not static inputs — they are a live feedback system.
Gamma exposure is not a scalar risk metric but a dynamic amplifier of hedging flow.
As spot moves, delta rebalancing generates secondary order flow which feeds back into execution conditions.
This creates a closed-loop system where hedging activity becomes part of price formation itself.
Any simulation that does not recompute Greeks continuously under realistic implied volatility dynamics
and does not simulate hedge execution latency is effectively modeling a different system than the one traded live.
Margin-aware execution
Real risk is not PnL — it is constrained liquidity under margin dynamics.
Portfolio margin constraints, intraday risk adjustments, and broker-level exposure limits
actively reshape execution feasibility. These constraints are non-linear and state-dependent.
A backtest that ignores margin compression effects will systematically overestimate survivability during stress regimes,
precisely when hedging demand and liquidity fragmentation are highest.
Transparency model
Forward execution is the only valid proof of system behavior.
For this reason, G-BOT relies exclusively on live or paper-traded forward sessions under real-time market conditions.
Every fill, hedge, rollover, and margin event is observable in real time on the operator console.
This preserves full state fidelity, including HTI evolution and exposure rebalancing logic.
Live streaming
Forward sessions on YouTube
Recorded operator sessions showing fills, hedging activity, drawdowns and recovery cycles without post-processing.
// dynamic delta hedging · live executiongreeks · margin · hedge flow · HTI state
// 05 · Library
Read the heretical bits before you decide.
The site is also a long-form essay collection. The pieces below are
where the operator voice comes through most clearly — they are the
best way to decide whether this initiative fits how you think about
markets.
G-BOT Algorithmic extends its core research into fundamental computational domains.
We are engineering high-performance architectures that transcend existing
bottlenecks in signal processing and data science.
Strategic IP Offering
Convolution Bid
A radical reformulation of convolution theory—the most ubiquitous operation in modern computing—applied to high-throughput domains including Radar, MRI, and AI inference.
By replacing legacy spectral methods with our proprietary state-propagation architecture, we deliver a computational efficiency breakthrough available for acquisition by a suitable strategic partner.
G-BOT is the practical realisation of decades of research by
Prof. Tom Gastaldi — author of the USE principle and a long-time
presence in algorithmic-trading research. He has been a guest
commentator on Italian prime-time television (Anno Zero,
Michele Santoro) and publishes both the academic and the
operator-facing pieces of this project himself.
arXiv · ResearchGate · Business Reporter · datatime.eu
// 07 · Next stage
From private initiative to publicly traded company.
This is not an offer or a solicitation. It is a statement of intent.
The long-term goal of G-BOT is to transition into a publicly traded
company in partnership with a major financial institution that has
substantial expertise in the IPO process. Early participants in the
private stage have a documented path into that transition.
We are actively in conversation with prospective Strategic Sponsors
— investors with the expertise and relationships to lead a
transition of this kind. If you recognise yourself in that
description, we invite a direct conversation.
Send a short note: confirm you are a professional investor, that
you understand the risks, and that you have (or will have) the
minimum funds in an IBKR Portfolio Margin account for the paper
evaluation phase.
You will receive the executable, setup notes, and a direct line
to a human operator. No forms. No funnel. No drip campaign.
