Price Action Pivoter™ V5Ai - Automated Trading System for NinjaTrader

Price Action Pivoter™ V5Ai (Build1006) – Unique Real-Time AI Trading App & AI Automated Trading Software Bot (LSTM Machine Learning) Built for NinjaTrader®, Featuring an Integrated Real-Time Gemini News AI Trading System. Developed by a Licensed & Registered Commodity Trading Advisor (CTA) — NFA #0516752

Eliminate Emotional Decision Fatigue — Trade with Systematic Edge

If you have ever frozen on a clean setup, second-guessed the market’s direction during a trade, or allowed an emotional override to derail an otherwise sound plan, Price Action Pivoter™ V5AiBuild1006 was created to address exactly that problem.

The system was developed by Raffi Sosikian, a software engineer, former money manager, Series 3 professional, and registered Commodity Trading Advisor with more than a decade of experience developing automated futures-trading systems.

God did not put us on this earth to spend our lives chained to computer monitors, following every bar like a never-ending tennis match while sacrificing our time, health, family, and peace of mind. We were meant to spend quality time with the people we love, enjoy the outdoors, and live our lives—not remain trapped in a chair, stressed and second-guessing every market movement.

Have you ever attended a social gathering and instinctively searched the room for another trader because traders seem to be the only people who truly understand the pressure? They understand the constant mental battle, the trades replaying in your head, and the way the market can consume your thoughts even when you are supposed to be relaxing.

That cycle needs to be broken.

Three Powerful Technologies Working as One

Price Action Pivoter™ V5AiBuild1006 combines three distinct technologies within one coordinated automated trading platform:

Price Action Pivoter™ is the NinjaTrader execution, money-management, trade-management, and risk-management engine.
HigherEdgeAI™ 3.0 is the Python-based machine-learning intelligence that evaluates historical relationships and produces qualifying directional signals.
PQNews™ 2.0 is the Gemini-powered real-time market-news intelligence that can block a LONG or SHORT direction when current news conflicts with the proposed trade.

HigherEdgeAI™ does not place trades directly, and PQNews™ does not originate trades. Price Action Pivoter™ receives their intelligence, applies its own price-action and eligibility requirements, and determines whether a qualifying trade should be executed and managed.

This is an important distinction. The system does not blindly enter whenever an AI score appears, nor does a favorable news assessment automatically create a trade. Every potential entry must still satisfy the strategy’s configured price-action, machine-learning, timing, money-management, and risk-management requirements.

Many automated systems depend on a largely fixed set of entry and exit rules. Those rules may respond well under certain market conditions but struggle when volatility, market structure, correlations, or the news environment changes.

Price Action Pivoter™ V5AiBuild1006 was designed differently. It combines current and historical price-action structure, more than 360 engineered machine-learning features, periodically retrained model candidates, related-market context, real-time news intelligence, and granular trade-management controls within one integrated decision process.

The objective is not to predict every market movement. It is to identify qualified opportunities, reject conditions that do not meet the configured standards, and execute the resulting decisions without hesitation, fatigue, revenge trading, or emotional interference.

Here’s How It Works

When you enable Price Action Pivoter™ V5AiBuild1006 on a qualifying futures chart, the strategy automatically launches HigherEdgeAI™ 3.0 and PQNews™ 2.0 in the background. You do not need to operate or enter commands into either Python terminal.

When the NinjaTrader strategy is disabled, the corresponding background applications automatically shut down.

Price Action Pivoter™ supplies HigherEdgeAI™ 3.0 with current and historical market information, including price-action structure, volatility, momentum, moving averages, Bollinger Band relationships, pivot levels, prior highs and lows, session structure, multiscale market behavior, and your selected AI Training Desired Profit and AI Training Risk Per Trade values.

HigherEdgeAI™ 3.0 transforms that information into more than 360 engineered features evaluated across rolling market sequences. Rather than relying on one permanently selected model path, HigherEdgeAI™ 3.0 trains and evaluates multiple competing machine-learning candidates, backtests their behavior, and selects the strongest qualified candidate for the available data.

Only the selected candidate’s qualifying signals are released to Price Action Pivoter™.

For supported equity-index instruments—including ES, MES, NQ, MNQ, YM, MYM, RTY, and M2K—the system can also incorporate contemporaneous directional context from related index markets. This helps identify whether correlated markets are confirming or conflicting with the primary instrument.

PQNews™ 2.0 further supports these markets through shared economic-event intelligence. An important economic-event update can be shared across the related index family without requiring every market instance to independently repeat the same analysis.

Machine Learning Built Around Your Objectives

What makes HigherEdgeAI™ 3.0 particularly different is that its training objective incorporates your selected profit and risk parameters.

You may enter your own AI Training Desired Profit and AI Training Risk Per Trade values, or you may use the built-in PinnacleQuantRangeTracker™ to obtain market-aligned recommendations based on recent trading ranges.

HigherEdgeAI™ 3.0 is designed to answer this question:

Based on the analyzed past data, does the model have sufficient confidence that, over the next 120 minutes, a trade in a particular direction will reach the selected AI Training Desired Profit before reaching the selected AI Training Risk Per Trade?

The model-confidence threshold is set to 80% by default and is user-configurable from 60% to 90% directly within the Price Action Pivoter™ strategy settings. When the calculated directional score meets or exceeds the selected threshold, HigherEdgeAI™ 3.0 may release a qualifying ConsiderLong or ConsiderShort signal; the selected threshold represents the model’s required confidence level, not a guaranteed win rate.

The powerful execution engine then takes over: it enters trades seamlessly based on the HigherEdgeAI™ 3.0 signals and pure price action. It dynamically manages every aspect of the trade — adjusting stops to reduce risk or lock in unrealized profits, scaling out, trailing positions with custom Peak High-Low logic, and even adapting profit targets and stops while the trade is live — all while strictly enforcing your overall risk and money management rules.

This unique automated trading system evaluates trading probability continuously, on every bar, using your own Desired Profit and Risk Per Trade parameters as the foundation. The result:

A probabilistic edge on every trade — engineered to extract maximum value from every trade.
No guessing.
No curve fitting.
No Renko bars or Heiken Ashi bars to mask results.

Everything runs hands-free, so you can finally step away from the screen and do something more meaningful with your time.

This is not another static rules-based bot, nor is it a black box algo. It’s a living, adaptive automated trading engine, which runs on a unique machine learning system that now incorporates real-time news (requires Gemini API key) that gets smarter with every session while staying fully under your control. It took a long time to get this unique automated trading system to where it is, and I’m grateful and humbled by all the clients that have been on this journey with me.

Real-Time News Intelligence Through PQNews™ 2.0

Historical data alone cannot know what was reported in the news moments ago, or the news that will come out since the trade’s entry.

Using your Gemini API key, which you will need to get separately, PQNews™ 2.0 evaluates current economic events, real-time news from sources all over the Internet, social media posts, market-moving headlines, geopolitical developments, and the market’s observed reaction. When the news environment materially conflicts with a proposed trade direction, PQNews™ can establish a directional veto that temporarily blocks LONG or SHORT participation, and then evaluates price action and price levels to determine if it’s safe to release that directional veto.

A veto does not create the opposite trade. It simply tells Price Action Pivoter™ not to participate in the blocked direction while the identified risk remains active.

Price action is then monitored to determine whether market behavior eventually justifies releasing that veto. This allows the strategy to respect both the headline and what the market actually does after the news becomes known.

Automated Execution and Granular Trade Management

Once a qualifying opportunity survives the strategy’s price-action requirements, HigherEdgeAI™ assessment, PQNews™ directional-veto checks, and all other enabled conditions, Price Action Pivoter™ can execute and manage the trade automatically.

Depending on your configuration, the system can coordinate:

Initial entries and qualifying scale-in entries
Profit targets and protective stop orders
Peak High-Low trailing-stop logic
Early Windfall protection
RRG-related trade management
Time-based and money-based exits
Adverse machine-learning signal exits
Dynamic risk reduction
Unrealized-profit protection
Session-loss controls
Session high-watermark profit protection
End-of-session trade management

These controls give you granular authority over how trades are entered, protected, adjusted, and exited. The system follows the parameters you select without becoming fearful after a loss, overconfident after a win, impatient during a quiet period, or tempted to take revenge on the market.

No automated system can eliminate trading risk, guarantee an order or fill, or ensure a particular trading result. Automation also does not mean that the platform, brokerage connection, orders, and account should be left completely unsupervised. You remain responsible for selecting appropriate settings and monitoring your trading environment.

What Price Action Pivoter™ can do is remove subjective, moment-to-moment decision-making from the automated process and replace it with a disciplined framework that consistently follows the configured rules.

The Result: More Intelligence. More Discipline. More Control.

Price Action Pivoter™ V5AiBuild1006 does far more than automate order placement. It brings machine learning, price action, correlated-market awareness, news intelligence, and advanced trade management together to give serious futures traders a more disciplined way to pursue their objectives.

Most importantly, you reclaim your time. Instead of spending every session staring at charts and wrestling with every market movement, you can allow a disciplined automated process to do the work it was configured to perform—giving you more freedom for your family, your business, and the life taking place beyond the screen.

Every trade must earn its way in. If the required machine-learning confidence and strategy conditions are absent, the system stands aside rather than forcing an unqualified trade.

Your trading rules remain rules. The automated engine does not become fearful after a loss, overconfident after a win, impatient during a quiet market, or tempted to revenge trade.

Qualified opportunities are acted upon without hesitation. When the enabled requirements align, the strategy can execute immediately according to your settings—without freezing, second-guessing, or waiting for emotional reassurance.

You can see the intelligence working. Model-confidence scores, directional signals, PQNews™ vetoes, event status, and strategy controls provide visibility into why the system is participating—or deliberately staying out.

The machine learning is built around your objectives. HigherEdgeAI™ 3.0 evaluates historical market relationships using your selected AI Training Desired Profit, AI Training Risk Per Trade, and user-configurable confidence threshold.

Related markets become actionable evidence. ES/MES, NQ/MNQ, YM/MYM, and RTY/M2K can incorporate contemporaneous context from related equity indexes to identify when correlated markets confirm—or conflict with—the primary instrument.

Breaking news becomes part of the decision process. PQNews™ 2.0 evaluates real-time headlines, economic events, geopolitical developments, and observed market reaction so that a conflicting LONG or SHORT direction can be blocked before the strategy participates.

You control the entire trade lifecycle. Configure entries, scale-ins, protective stops, profit targets, Peak High-Low trailing, Early Windfall, RRG management, time and money exits, adverse-AI exits, session-loss controls, and session high-watermark protection.

You gain automation without surrendering control. Your settings govern the strategy, while you retain the authority to enable, disable, supervise, or adjust the system as your objectives change.

This is not another one-dimensional entry bot. It is an integrated automated trading platform combining price action, machine learning, real-time news intelligence, execution, and advanced money, trade, and risk management.

The underlying technology remains proprietary, but you retain control over the settings that govern its behavior in this self-directed system. You determine the operating windows, training objectives, confidence threshold, trade-management features, and risk parameters used by the strategy.

Built for Serious Futures Traders

If the system you currently use is working well for you, you do not need to replace it. Keep it—and may God bless it even more.

But if you are being honest with yourself and your existing approach is not providing the discipline, adaptability, control, or freedom you have been seeking, continuing to repeat the same process may cost you considerably more in lost time, emotional energy, and trading capital.

Pinnacle Quant is a registered NinjaTrader Ecosystem vendor, and I have spent more than a decade developing algorithmic futures-trading systems for self-directed and professional traders. Other trading-software vendors have also engaged me to help solve difficult strategy-development problems within their own automated systems.

That experience is reflected throughout Price Action Pivoter™ V5AiBuild1006—from its entry logic and machine-learning integration to its protective-stop engine, session controls, news intelligence, and highly granular configuration.

If you have limited excess capital or are choosing trading software primarily according to the lowest available price, this system will probably not be the right fit—and that is perfectly okay. I sincerely wish you success.

However, if you have sufficient risk capital and are looking for a serious, professionally engineered automated trading platform—one that provides advanced intelligence without taking control away from you—this is where the difference becomes clear.

Price Action Pivoter™ V5AiBuild1006 gives you access to an integrated decision and execution platform powered by HigherEdgeAI™ 3.0 and PQNews™ 2.0. It evaluates historical relationships, current price action, related-market behavior, and real-time news before automatically managing qualifying trades according to your configured objectives.

Replace hesitation with rules.

Replace emotional overrides with disciplined execution.

Replace endless screen-watching with technology designed to give you more control over your trading—and more freedom over your time.

Adaptive Calibration—Aligned With What the Market Is Actually Offering

Stop forcing the same static profit and risk targets onto a market that never remains exactly the same.

Volatility changes from one session to the next. A profit objective that is reasonable during an active, wide-ranging session may be unrealistic during quiet, low-volume conditions. Conversely, parameters designed for a quiet market may fail to reflect the opportunity available when market ranges expand.

Price Action Pivoter™ V5AiBuild1006 addresses this problem through the built-in PinnacleQuantRangeTracker™.

PinnacleQuantRangeTracker™ analyzes recent market behavior by separating Regular Trading Hours and Extended Trading Hours, examining two-hour session segments, and evaluating daily ranges. It provides a data-driven reference for selecting AI Training Desired Profit and AI Training Risk Per Trade values that better reflect the ranges the instrument has recently produced.

The tool does not predict tomorrow’s range or guarantee that a recommended objective will be reached. It gives you a statistically grounded alternative to selecting training values arbitrarily.

How PinnacleQuantRangeTracker™ Works

Before you enable the strategy, PinnacleQuantRangeTracker™ can analyze recent session data and:

Separate Regular Trading Hours (RTH) and Extended Trading Hours (ETH)
Divide each session into two-hour windows
Calculate the maximum, minimum, and median ranges produced during those windows
Evaluate recent daily ranges
Apply anomaly filtering intended to prevent unusual outlier sessions from distorting the recommendations
Produce market-aligned reference values for AI Training Desired Profit and AI Training Risk Per Trade

You remain in control through two operating choices:

Manual Mode: Enter your own AI Training Desired Profit and AI Training Risk Per Trade values according to your objectives and risk tolerance.
Dynamic Mode: Enable Use PQRangeTracker™ AI Training and allow the strategy to calculate and apply market-aligned values using recent RTH, ETH, and daily range behavior (see an example below).

In either mode, your chosen AI Training parameters or choosing PQRangerTracker™ to choose the AI training parameters, HigherEdgeAI™ 3.0 uses the selected values as part of its machine-learning objective. The model is therefore calibrated not only around what you want the trade to accomplish, but also around what the market has recently demonstrated that it is capable of producing.

That is professional-level adaptability designed for serious, self-directed futures traders—while keeping the final operating decisions firmly under your control.

Engineering Philosophy: The Foundation of Price Action Pivoter™ V5Ai

When you look under the hood of Price Action Pivoter™ V5Ai, you aren’t just seeing code — you’re seeing the culmination of more than a decade of iterative development, real-market testing, and architectural refinement by a licensed and registered Commodity Trading Advisor (CTA) and professional algorithmic systems engineer.

My design philosophy rests on three core pillars that separate a professional-grade execution engine from typical trading bots:

1. Deterministic Execution vs. Heuristic Guesswork

Most automated trading systems depend heavily on lagging indicators that can introduce decision delays. Price Action Pivoter™ V5Ai reduces that reliance through Pinnacle Quant’s proprietary Python-based multivariate LSTM model, HigherEdgeAI™ 3.0. This machine-learning model analyzes historical data and evaluates 360 engineered features per market bar across rolling 30-bar sequences, producing 10,800 time-structured numerical inputs for every model decision. It generates forward-looking directional confidence scores in real time before a trade can qualify, rather than merely reacting after a market move is already underway.

2. Probabilistic Forecasting & Adaptive Calibration

The machine-learning system in HigherEdgeAI™ 3.0 is designed to answer one central question:

Based on the analyzed past data, does the model have sufficient confidence that, over the next 120 minutes, a trade in a particular direction will reach the selected AI Training Desired Profit before reaching the selected AI Training Risk Per Trade?

The required model-confidence threshold is set to 80% by default and is user-configurable from 60% to 90% within the Price Action Pivoter™ strategy settings. This threshold is the minimum confidence required for a qualifying directional signal—not a guaranteed win rate—and no directional signal is released when the selected threshold is not met.

When the optional PinnacleQuantRangeTracker™ is enabled within Price Action Pivoter™ V5AiBuild1006, it analyzes recent two-hour RTH and ETH ranges, along with daily market ranges, to calculate market-aligned AI Training Desired Profit and AI Training Risk Per Trade values. HigherEdgeAI™ 3.0 then uses these dynamically calculated objectives while training and evaluating its competing machine-learning candidates.

This creates genuinely range-adaptive behavior: rather than forcing the same static expectations onto every market environment, the system recalibrates its machine-learning objectives around the ranges the instrument has recently demonstrated. The market does not bend to a trader’s expectations—so this unique trading engine adapts its expectations to what the market is actually offering.

3. Asymmetric Information Processing + Granular Execution Intelligence

Speed matters in mostly algo run futures markets—but speed without context can simply accelerate a bad decision. The greater advantage comes from recognizing which information deserves weight before the strategy commits capital.

The integrated PQNews™ 2.0 intelligence layer, powered through the client’s Gemini API key, does more than identify that news exists. It evaluates real-time headlines, economic events, geopolitical developments, and the market’s observed reaction to determine whether current information supports participation, warrants caution, or conflicts strongly enough to establish a LONG or SHORT directional veto. PQNews™ does not originate trades; it provides current-event intelligence that Price Action Pivoter™ evaluates alongside HigherEdgeAI™ signals, price action, related-market context, and every other enabled entry requirement.

Once a trade qualifies, Price Action Pivoter™ V5AiBuild1006 takes over through its sophisticated, multi-layer trade and capital-management engine. Configurable multi-leg scaling, dynamic Peak High-Low trailing stops, Near-Exit Profit Lock, Rapid Rejection Guard, Stop Competition, Early Windfall, adverse-AI and time-based exits, and precise per-trade money-management controls work together to pursue favorable market movement while applying structured protection as conditions change.

Above the individual-trade controls is the optional Session High-Watermark Profit Protection. Once session profit reaches your configured dollar trigger, the system tracks the highest session profit achieved and applies your selected percentage-based backstop floor—helping prevent a strong trading session from surrendering an unacceptable portion of its gains as market conditions reverse.

Together, these layers allow the system to evaluate news before participating, manage each qualifying trade dynamically, and protect performance at both the individual-trade and full-session levels—all according to the parameters you control.

See Backtest Results for Price Action Pivoter™ V5AiBuild1006 – Evidence for Traders Comparing the Best Automated Futures Trading Systems

Each backtest below includes:

NinjaTrader Strategy Analyzer Summary, monthly, and yearly performance breakdowns
Complete dashboard showing all automated trading strategy parameters
PDF printout of the complete backtest dashboard with precise parameters used in the strategy
The actual trade signals generated by HigherEdgeAI™ 3.0
Spreadsheet detailing every trade (entry price, exit price, profit/loss, etc.)

These backtests were not curve-fitted. The core parameters remained consistent across the entire 32-month period, demonstrating the robustness of this real-time machine learning driven automated trading system built exclusively for the NinjaTrader platform.

Some notes on the backtests:

Period: January 1, 2024 – August 24, 2026 (32 months)
HigherEdgeAI™ 3.0 Confidence Threshold: 80% and 90% was used for both ES and NQ.
Schedule: Monday through Friday during Regular Trading Hours (RTH), including days with major news, geopolitical events, or market shocks
Bar Type: Standard 1-minute bars only calculated on bar close
No Renko, Heiken-Ashi, or custom bars as NinjaTrader Support states Renko bars, Heiken-Ashi and other such bars are not reliable in backtests. Only standard minute bars produce accurate results on backtests, and only if strategy is calculated on bar close
Market Data: Uses live market data feed (Kinetick, Rithmic, Continuum, CQG, etc.). Does not rely on tick data
Machine Learning Training for HigherEdgeAI™ 3.0: AI Training Desired Profit & AI Training Risk Per Trade were automatically picked by the system via PinnacleQuantRangeTracker™
Contract Quantity: each backtest demonstrates 1 contract and 2 contracts, plus any additional scale-in legs.
Slippage: 2 ticks included in all results
Commissions: Not included — multiply the total number of trades by your broker’s round-trip commission and deduct from the reported net profit

ES – Quantity: 1 Contract, 80% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 1 – Instrument: ES ($813,162.50)
Net Profit: $813,162.50
Gross Profit: $967,800.00
Gross Loss: ($154,637.50)
Profit Factor: 6.26
Max Drawdown: ($5,850.00)
Number of Trades: 2,421
Win Rate: 82.94%
Number of Winning Trades: 2,008
Number of Losing Trades: 366
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.05
Average Time in Market: 67.23 min
Max Time to Recover: 28.05 days
Longest Flat Period: 10.77 days
Average MAE: $349.14
Average MFE: $619.36
Market Condition: All including Economic News Events
Contract(s): 1
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $5 X 2,421 = $12,105 paid in commission.

ES – Quantity: 2 Contracts, 80% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 2 – Instrument: ES ($1,626,325.00)
Net Profit: $1,626,325.00
Gross Profit: $1,935,600.00
Gross Loss: ($309,275.00)
Profit Factor: 6.26
Max Drawdown: ($11,700.00)
Number of Trades: 2,421
Win Rate: 82.94%
Number of Winning Trades: 2,008
Number of Losing Trades: 366
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.05
Average Time in Market: 67.23 min
Max Time to Recover: 28.05 days
Longest Flat Period: 10.77 days
Average MAE: $698.28
Average MFE: $1,238.72
Market Condition: All including Economic News Events
Contract(s): 2
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $10 (2 contracts * $5) X 2,421 = $24,210 paid in commission.

ES – Quantity: 1 Contract, 90% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 3 – Instrument: ES ($812,075.00)
Net Profit: $812,075.00
Gross Profit: $958,200.00
Gross Loss: ($146,125.00)
Profit Factor: 6.56
Max Drawdown: ($3,587.50)
Number of Trades: 2,358
Win Rate: 83.55%
Number of Winning Trades: 1,970
Number of Losing Trades: 344
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.12
Average Time in Market: 64.76 min
Max Time to Recover: 18.83 days
Longest Flat Period: 10.89 days
Average MAE: $340.22
Average MFE: $625.56
Market Condition: All including Economic News Events
Contract(s): 1
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $5 X 2,358 = $11,790 paid in commission.

ES – Quantity: 2 Contracts, 90% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 4 – Instrument: ES ($1,624,150.00)
Net Profit: $1,624,150.00
Gross Profit: $1,916,400.00
Gross Loss: ($292,250.00)
Profit Factor: 6.56
Max Drawdown: ($7,175.00)
Number of Trades: 2,358
Win Rate: 83.55%
Number of Winning Trades: 1,970
Number of Losing Trades: 344
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.12
Average Time in Market: 64.76 min
Max Time to Recover: 18.83 days
Longest Flat Period: 10.89 days
Average MAE: $680.45
Average MFE: $1,251.11
Market Condition: All including Economic News Events
Contract(s): 2
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $10 (2 contracts * $5) X 2,358 = $23,580 paid in commission.

NQ – Quantity: 1 Contract, 80% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 5 – Instrument: NQ ($1,218,115.00)
Net Profit: $1,218,115.00
Gross Profit: $1,521,740.00
Gross Loss: ($303,625.00)
Profit Factor: 5.01
Max Drawdown: ($8,100.00)
Number of Trades: 3,080
Win Rate: 85.03%
Number of Winning Trades: 2,619
Number of Losing Trades: 353
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.34
Average Time in Market: 29.30 min
Max Time to Recover: 14.03 days
Longest Flat Period: 10.76 days
Average MAE: $470.31
Average MFE: $735.11
Market Condition: All including Economic News Events
Contract(s): 1
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $5 X 3,080 = $15,400 paid in commission.

NQ – Quantity: 2 Contracts, 80% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 6 – Instrument: NQ ($2,436,230.00)
Net Profit: $2,436,230.00
Gross Profit: $3,043,480.00
Gross Loss: ($607,250.00)
Profit Factor: 5.01
Max Drawdown: ($16,200.00)
Number of Trades: 3,080
Win Rate: 85.03%
Number of Winning Trades: 2,619
Number of Losing Trades: 353
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.34
Average Time in Market: 29.30 min
Max Time to Recover: 14.03 days
Longest Flat Period: 10.76 days
Average MAE: $940.62
Average MFE: $1,470.21
Market Condition: All including Economic News Events
Contract(s): 2
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $10 X 3,080 = $30,800 paid in commission.

NQ – Quantity: 1 Contract, 90% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 7 – Instrument: NQ ($1,217,730.00)
Net Profit: $1,217,730.00
Gross Profit: $1,474,325.00
Gross Loss: ($256,595.00)
Profit Factor: 5.75
Max Drawdown: ($4,700.00)
Number of Trades: 3,003
Win Rate: 85.11%
Number of Winning Trades: 2,556
Number of Losing Trades: 321
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.25
Average Time in Market: 30.56 min
Max Time to Recover: 6.82 days
Longest Flat Period: 10.77 days
Average MAE: $461.55
Average MFE: $735.79
Market Condition: All including Economic News Events
Contract(s): 1
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $5 X 3,003 = $15,015 paid in commission.

NQ – Quantity: 2 Contracts, 90% HigherEdgeAI™ 3.0 Confidence Threshold

Backtest 8 – Instrument: NQ ($2,435,460.00)
Net Profit: $2,435,460.00
Gross Profit: $2,948,650.00
Gross Loss: ($513,190.00)
Profit Factor: 5.75
Max Drawdown: ($9,400.00)
Number of Trades: 3,003
Win Rate: 85.11%
Number of Winning Trades: 2,556
Number of Losing Trades: 321
Start Date: 1/1/2024
End Date: 8/24/2026
Hours Traded: 6:30AM to 12:30PM Pacific Time
Flatten Everything: 12:59PM Pacific Time
Sortino Ratio: 1.00
Sharpe Ratio: 2.25
Average Time in Market: 30.56 min
Max Time to Recover: 6.82 days
Longest Flat Period: 10.77 days
Average MAE: $923.10
Average MFE: $1,471.58
Market Condition: All including Economic News Events
Contract(s): 2
Slippage: 2
Commission: Must calculate by multiplying number of trades by round trip commission. For example, $10 (2 contracts * $5) X 3,003 = $30,030 paid in commission.

While no backtest is 100% accurate and past results are not necessarily indicative of future results, the above hypothetical backtests provide valuable insights into how this innovative machine learning automated trading system uses real-time A.I. signals to identify trades and manage them throughout their full lifecycle, with granular user control over entries, exits, and risk management.

Best Practice Recommendation

For optimal performance in this day-trading automated trading system, I recommend re-enabling Price Action Pivoter™ V5Ai each trading day.

Here’s the simple routine:

In NinjaTrader’s Data Series, set your chart to 5 Days to Load.
Before bed, disable the strategy.
Schedule HigherEdgeAI™ 3.0 to run approximately one hour before your initial trade time
If you’re trading more than one instrument, stagger each instrument’s scheduling time by 8 minutes. For example, assuming your Trade Start Time 1 is at 830, then schedule HigherEdgeAI™ 3.0 to run at 730 for instrument one, 738 for instrument two, 746 for instrument three and so on.

API Key Requirement

The system requires an API key, assigned exclusively to active license holders. This key connects to our server to retrieve the algorithms necessary for the model. The design ensures that critical code is stored securely on the server and not included in the software itself. Each client receives a unique API key that connects to our secure server, where the machine learning algorithm is hosted and dynamically maintained. This ensures the model remains centralized, and protected, delivering signals that respond to evolving market conditions.

Trade Nearly Any Futures Instrument with Real-Time Machine Learning Automation

Price Action Pivoter™ V5Ai (Build 1006) is designed to work on virtually any futures instrument, including:

$ES, $NQ, $GC, $SI, $FDAX, $YM, $RTY, $CL, $MES, $MNQ, and many more.

As long as the instrument provides at least 22.5 hours of trading data per session (covering both RTH and ETH), the system can effectively analyze price action and generate signals using its real-time HigherEdgeAI™ 3.0 multivariate LSTM machine learning model.

This flexibility makes Price Action Pivoter™ V5Ai one of the most versatile automated trading systems available on the NinjaTrader 8 platform.

How I Integrated Python with NinjaTrader — and Why I Didn’t Use .NET-Based Python Libraries
Rather than rely on less-maintained .NET-based Python libraries with limited community support and slower update cycles, I chose to build a robust integration that lets Python and NinjaTrader work together seamlessly—each doing what it does best.

Python handles the machine learning (via the proprietary HigherEdgeAI™ 3.0 multivariate LSTM machine learning model), signal generation, and data processing, while NinjaTrader serves as the trade execution and trade management engine—managing entries, exits, and risk in real time based on those signals.

This architecture enables faster iteration, more advanced model development, and greater transparency—since Python offers access to cutting-edge AI frameworks, well-maintained libraries, and a significantly broader developer community than .NET.

Fully Automated Real-Time AI (Machine Learning) Trading System:
This is a self-directed fully automated trading system designed for futures traders who are comfortable with market risk and serious about long-term performance. It’s built for those who seek to deploy a sophisticated, real-time AI-driven strategy they can rely on for years to come—traders who understand the importance of operating with a probabilistic edge on every trade.

HigherEdgeAI™ 3.0 Multivariate LSTM Machine Learning Python Terminal:

HigherEdgeAI™ 3.0 is Pinnacle Quant’s most advanced Python-based real-time multivariate LSTM machine-learning system. It builds upon the original HigherEdgeAI™ architecture by evaluating more than 360 engineered features, training and comparing multiple model candidates, and releasing only the selected model’s qualifying directional signals to Price Action Pivoter™ V5AiBuild1006.

HigherEdgeAI™ 3.0 is engineered to use the computer’s available processing resources efficiently so that its data preparation, model evaluation, retraining, and real-time signal analysis can operate alongside NinjaTrader. Unlike standalone machine-learning applications, HigherEdgeAI™ 3.0 is launched and controlled automatically by Price Action Pivoter™. When you enable the strategy, HigherEdgeAI™ 3.0—and PQNews™ 2.0 when configured—starts automatically; when you disable the strategy, the corresponding terminals automatically shut down. You do not need to type commands into, minimize, close, or otherwise operate the Python terminals yourself.

A stable, wired Internet connection is required for live automated trading; do not rely on Wi-Fi. If the Internet connection is interrupted, immediately verify all account positions and working orders through the NinjaTrader Control Center and manually close or cancel anything that remains active. Once the connection has been restored and you have confirmed that the account is flat with no working orders, close the affected chart, open a new chart, and re-enable Price Action Pivoter™ V5AiBuild1006.

Here is what you will see in the Python terminal when it first loads HigherEdgeAI™ 3.0:

Here is what you will see when the HigherEdgeAI™ 3.0 Multivariate LSTM machine learning model is running:

What’s New in Price Action Pivoter™ V5AiBuild1006

Price Action Pivoter™ V5AiBuild1006 is a major intelligence, trade-management, and session-protection upgrade—not merely a news-system update. Build1006 introduces HigherEdgeAI™ 3.0, the new PQNews™ 2.0 intelligence layer, user-configurable machine-learning confidence, improved PinnacleQuantRangeTracker™ training-objective selection, granular protective-stop management for every scale-in leg, and the new Session High-Watermark Profit Lock.

HigherEdgeAI™ 3.0—More Data, Competing Models and User-Controlled Confidence

HigherEdgeAI™ 3.0 replaces HigherEdgeAI™ 2.0 with a substantially expanded machine-learning architecture. It analyzes historical data and evaluates 360 engineered features per market bar across rolling 30-bar sequences, producing 10,800 time-structured numerical inputs for every model decision. It trains and compares multiple competing model candidates and releases only the selected model’s qualifying directional signals to Price Action Pivoter™.

HigherEdgeAI™ 3.0 is designed to answer this question:

Based on the analyzed past data, does the model have sufficient confidence that, over the next 120 minutes, a trade in a particular direction will reach the selected AI Training Desired Profit before reaching the selected AI Training Risk Per Trade?

Build1006 gives traders direct control over what qualifies as “sufficient confidence.” The HigherEdgeAI™ 3.0 Signal Threshold is set to 80% by default and is user-configurable from 60% to 90% within the Price Action Pivoter™ strategy settings.

When the calculated directional score meets or exceeds the selected threshold, HigherEdgeAI™ 3.0 may release a qualifying ConsiderLong or ConsiderShort signal. When the threshold is not met, no directional signal is released—the model is never forced to manufacture a trade. The selected percentage is a model-confidence requirement, not a guaranteed win rate.

PQNews™ 2.0—Live News Scoring, Directional Vetoes and Main Headline Watch

Price Action Pivoter™ V5Ai Build1006 introduces PQNews™ 2.0, a Gemini-powered, live-web-search intelligence layer that evaluates current financial reporting, headline news from sources such as CNBC, Fox Business, and Reuters, economic releases, official policy communications, geopolitical developments, corporate announcements, and observed market reactions.

PQNews™ 2.0 produces a News Score from 0.0 to 10.0, classifies the current news environment as HIGH or LOW impact, and independently evaluates whether a specific catalyst creates sufficient one-sided risk to establish a LONG or SHORT Directional Veto.

Its separate Main Headline Watch checks lead headlines on CNBC, Fox Business, and Reuters during configured trading windows. When a qualifying headline reports that a relevant market is sharply rising or falling, PQNews™ 2.0 can establish the corresponding directional trade veto without waiting for the next comprehensive news-analysis cycle. As illustrated in the screenshot above, the system automatically monitors price action in real time to determine whether and when the directional veto may safely be released.

PQNews™ never originates a trade or overrides protective stops and profit targets. Price Action Pivoter™ evaluates its news intelligence alongside HigherEdgeAI™ 3.0 signals, price action, related-market context, and every other enabled requirement.

Session High-Watermark Profit Lock—Protect the Trading Day, Not Merely One Position

Build1006 introduces the optional Session High-Watermark Profit Lock, extending protection from individual trades to the cumulative performance of the entire trading session.

Once session net profit reaches the user-configured High-Watermark Session Profit Trigger $, the feature arms and begins tracking the highest session profit achieved. It then calculates a protected floor using the selected High-Watermark Session Profit Backstop Floor %.

As session profit reaches new highs, the protected floor ratchets upward with it—it never moves downward. The system applies that floor through protective stops on the open trade legs, while permitting another stop system to remain in control whenever it provides tighter protection.

If the High-Watermark floor ultimately closes the position, Price Action Pivoter™ flattens any remaining contracts and blocks additional entries, reversals, and scale-ins for the remainder of the session. This helps prevent a strong trading day from surrendering an unacceptable portion of its accumulated gains.

Granular Scale-In Protection—Independent Stops for L1, L2 and L3

Build1006 substantially strengthens the management of scale-in positions. Rather than protecting all added contracts through one blended position-level calculation, the strategy can track and manage Scale-In L1, L2, and L3 independently.

Each scale-in leg can have its own:

Fill-price-based protective stop
Scaling MyRiskPerTrade distance
Peak-unrealized-profit activation threshold
HigherEdgeAI™-assisted alternate activation threshold
Peak High-Low trailing-stop calculation
Custom trailing percentage
Temporary scale-in profit target
Profit-target bar window
Near-Exit Profit Lock evaluation
Remaining quantity and confirmed protective-stop state

This provides far more precise protection than treating every added contract as one indistinguishable combined position. A profitable scale-in leg can tighten independently, while another leg continues operating according to its own fill price, profit history, and protective requirements.

Build1006’s shared protective-stop engine also compares competing stop candidates and applies the most protective valid stop for each affected leg. A different stop system cannot replace already confirmed protection with a looser stop, except for narrowly defined same-feature reversion behavior—and even those reversions cannot displace tighter protection established by another stop owner.

Improved PinnacleQuantRangeTracker™ AI Training Selection

Build1006 includes a substantially refined PinnacleQuantRangeTracker™ process for selecting AI Training Desired Profit and AI Training Risk Per Trade.

When Use PQRangeTracker AI Training is enabled, the system evaluates completed regular-session and overnight-session ranges, recent RTH expansion or contraction, overnight-range behavior, and the direction of the recent range trend. It uses those completed-session measurements to calculate a market-aligned AI Training Desired Profit rather than relying on an arbitrary fixed number.

The resulting Desired Profit is constrained within defined percentages of the recent completed-session baseline, preventing unusually small or extreme values from dominating the training objective. AI Training Risk Per Trade is then calculated at 75% of the resolved AI Training Desired Profit, rounded up to the next whole dollar.

Because only completed sessions are used, the selected values remain consistent regardless of what time Price Action Pivoter™ is enabled. Traders who prefer complete manual control may leave Dynamic Mode disabled and enter their own AI Training Desired Profit and AI Training Risk Per Trade values.

One Coordinated V5AiBuild1006 Intelligence and Protection System

Together, these Build1006 upgrades create a more advanced and coordinated trading platform:

HigherEdgeAI™ 3.0 evaluates historical relationships and produces qualifying directional intelligence.
The user-configurable 60%–90% threshold determines how much model confidence is required.
PinnacleQuantRangeTracker™ provides market-aligned profit and risk objectives for machine-learning training.
PQNews™ 2.0 evaluates current news and can block participation when real-time information conflicts.
Granular scale-in protection independently manages every L1, L2, and L3 entry.
Session High-Watermark Profit Protection helps defend accumulated performance across the entire trading session.
Price Action Pivoter™ V5AiBuild1006 coordinates the final entry, execution, money-management, trade-management, and risk-management decisions.

Build1006 therefore improves not only how the system identifies potential opportunities, but also how selectively it participates, how precisely it protects every entry leg, and how aggressively it defends the cumulative performance of the trading session.

What Do I Need to Get Started with Price Action Pivoter™ V5Ai (Build 1006)

Getting started is straightforward. You will need the following:

NinjaTrader 8 Platform
A brokerage account or prop firm account supported by NinjaTrader
Live market data feed (Kinetick, Rithmic, Continuum, CQG, or NinjaTrader Brokerage)
Stable wired internet connection (WiFi is not recommended for reliable performance)
Excess trading capital appropriate for futures trading (check margin requirements with your broker)

Hardware & System Requirements

Because Price Action Pivoter™ V5Ai runs a dedicated Python terminal via HigherEdgeAI™ 3.0 multivariate LSTM machine learning model for each instrument and works with PQNews™ 2.0 (which requires a Gemini API key) and runs its own Python terminal for each instrument, I recommend the following minimum specifications:

Memory (RAM):

Base requirement: 8 GB for NinjaTrader 8
+ 8 GB for each instrument traded concurrently (due to dedicated Python terminals)
+ 8 GB extra recommended for PQNews™ Checker AI

Examples:

Trading 1 instrument (e.g., ES only) → Minimum 24 GB RAM recommended
Trading 2 instruments (e.g., ES + NQ) → Minimum 32 GB RAM recommended

Storage: Minimum 10 GB free on an SSD drive CPU: Dual-core or better, ideally a modern quad-core or higher (Intel i7 / AMD Ryzen 7 or equivalent) with clock speed of 3.0 GHz+

Pro Tip: Approximately 70% of our clients run the system on a VPS rented from NinjaTrader Ecosystem-approved hosting providers. This provides excellent stability and performance. About 25% use desktop machines, and 5% run their own Windows servers.

If you’d like help choosing the right VPS or hardware setup, feel free to ask during your Zoom onboarding call with Raffi.

Engineered for Active Traders Who Seek a Higher-Probability Edge

If you’re an active futures trader with real capital at work — and you’re not looking for gimmicks, “green arrow” signals, or curve-fitted nonsense — then Price Action Pivoter™ V5Ai with HigherEdgeAI™ 3.0 real-time multivariate LSTM machine learning was built for you. I have been blessed to serve hundreds of active futures day traders and investors from all walks of life, and I look forward to serving you too!

Watch Live Trades with Price Action Pivoter™ V5AiBuild1006 with HigherEdgeAI™ 3.0 PQNews™ 2.0 Enabled

Watch Live Trades with Price Action Pivoter™ V5AiBuild1006 with HigherEdgeAI™ 3.0 & PQNews™ 2.0 Enabled

Watch the Last Hour of RTH trading LIVE using Price Action Pivoter™ V5AiBuild1006 with PQNews™ 2.0 Enabled

Watch how Price Action Pivoter™ V5Ai Build 1005, PQNews™, and HigherEdgeAI™ 2.0 established the foundational mechanics powering our upgraded Build 1006, PQNews™ 2.0, and HigherEdgeAI™ 3.0 releases.

Watch This Video To See Exactly How Price Action Pivoter™ V5Ai (Original Version – not V5AiBuild1005) Trains, Analyzes and Executes Trades:

UPDATE: March 21st, 2025 – NinjaTrader Joins Kraken & The Email Sent by CEO of NinjaTrader Martin Franchi:

Watch this interview Matthias Knab conducted with me on Opalesque TV during the SALT Conference in 2019, where we discussed my trading methodology and my work at the time as an E-mini S&P 500 (ES) specialist serving hedge fund portfolio managers. The interview pre-dates Pinnacle Quant’s retail focus, but the institutional foundation it reflects is what grounds every automated trading system I’ve built since:

Version History and Release Notes

Current Version—V5AiBuild1006 (Released: August 2026)

HigherEdgeAI™ 3.0.5.0: Adds 18 engineered pattern variables (378 engineered variables in total), enhanced data validation, and stronger model compatibility checks.
HigherEdgeAI™ 3.0.3.0: Replaces HigherEdgeAI™ 2.0 with a substantially expanded machine-learning architecture that evaluates 360 engineered features per market bar across rolling 30-bar sequences, producing 10,800 time-structured numerical inputs for every model decision. Version 3.0.3.0 expanded the model from 264 features and 7,920 inputs per sequence to 360 features and 10,800 inputs per sequence. This adds 96 per-bar variables and 2,880 time-structured inputs to every model decision. HigherEdgeAI™ trains and compares multiple competing model candidates and releases only the selected model’s qualifying directional signals to Price Action Pivoter™.
User-Configurable HigherEdgeAI™ Confidence Threshold: The minimum qualifying model-confidence threshold is set to 80% by default and can be adjusted from 60% to 90% directly within the Price Action Pivoter™ strategy settings. When the selected threshold is not met, no qualifying directional signal is released. The threshold is a model-confidence requirement, not a guaranteed win rate.
Cross-Index Machine-Learning Context: Supported equity-index instruments—including ES, MES, NQ, MNQ, YM, MYM, RTY, and M2K—can incorporate contemporaneous directional context from related index markets. This helps HigherEdgeAI™ recognize when correlated markets confirm or conflict with the primary instrument.
PQNews™ 2.0.4.2 (Released: October 2026): Every headline is confirmed from a live Google search result or the publisher’s own news feed, and its publication time is checked, before it can establish a Directional Veto. A separate publisher-feed check reads the Fox Business, CNBC, and Bureau of Labor Statistics feeds every 60 seconds without using the Gemini API key. Immediate economic-news research runs within the configured trading windows when a scheduled economic release is due, when price makes an unusually large move around a release, or when a release has produced no headline coverage; speeches and press conferences are tracked as scheduled remarks rather than data releases.
PQNews™ 2.0.4.1: Replaces the original PQNews™ Checker AI with a substantially expanded Gemini-powered intelligence layer. Version 2.0.4.1 strengthens counter-trend detection so PQNews™ can identify and veto conflicting trade directions more effectively. It also significantly reduces Gemini API usage while preserving real-time news analysis. PQNews™ continues to provide a News Score from 0.0 to 10.0, HIGH or LOW news-impact classification, independent LONG and SHORT Directional Vetoes, economic-event awareness, and recorded session history for forward-from-live backtesting.
Main Headline Watch: Adds a separate live-web-search process that checks the current lead headlines on CNBC, Fox Business, and Reuters during configured trading windows. A qualifying headline can establish a Directional Veto without waiting for the next comprehensive PQNews™ analysis cycle.
Price-Confirmed Veto Release: After at least 15 new one-minute bars have printed, releasing a LONG veto requires two consecutive closes above the closing price recorded when the veto began, with each close also higher than the close five bars earlier, followed by a close above the nearest monitored resistance level that was above price when the veto began. Releasing a SHORT veto requires the opposite: two consecutive closes below the closing price recorded when the veto began, with each close also lower than the close five bars earlier, followed by a close below the nearest monitored support level that was below price when the veto began. Once these price-action and price-level confirmations are satisfied, PQNews™ releases the veto; when no qualifying support or resistance level is available, the veto remains active through the end of the trading session.
Shared Economic-Event and Headline Intelligence: Charts within the same market group can share qualifying news and headline findings. This reduces duplicated Gemini requests and potential API usage while helping related instruments respond consistently to the same market-moving event.
Session High-Watermark Profit Lock: Introduces session-level profit protection above the individual-trade controls. Once session profit reaches the configured dollar trigger, the strategy tracks the highest session profit achieved and ratchets a protected percentage-based floor upward as the session high-watermark increases.
High-Watermark Session Lockout: If the Session High-Watermark floor ultimately closes the position, the strategy flattens any remaining contracts and blocks additional entries, reversals, and scale-ins for the rest of the session. If another ordinary trade exit closes the position before the High-Watermark protection executes, trading may continue normally.
Granular Scale-In Protective Stops: Strengthens the independent management of Scale-In L1, L2, and L3. Each scale-in leg maintains its own fill-price-based protective stop, remaining quantity, desired stop price, pending stop request, confirmed stop price, trailing thresholds, and protective-stop ownership.
Independent Scale-In Trailing Controls: Each L1, L2, and L3 scale-in leg can use its own peak-unrealized-profit activation threshold, HigherEdgeAI™-assisted alternate threshold, trailing percentage, protective-stop calculation, custom profit target, and profit-target bar window.
Full-Quantity Protection Confirmation: Build1006 verifies that protective-stop coverage is confirmed for the full remaining quantity of each tracked entry leg. When full protection cannot be maintained, the strategy’s protective fail-safe logic can request the appropriate market exit.
Reinforced Stop Competition: Multiple protective systems can run simultaneously, including MyRiskPerTrade, Rapid Rejection Guard, Peak High-Low trailing, MidPivotPointsPHL™, M-ATR, Near-Exit Profit Lock, Early Windfall Profit-Lock, and the Session High-Watermark floor. The most protective valid stop controls, and one stop system cannot loosen tighter protection already confirmed by another.
Protective Stop Order Safety Buffer: Adds a configurable dollar-based safety cushion for ordinary protective-stop placement. The amount is converted into ticks using the instrument’s one-contract PointValue and TickSize and is subject to the strategy’s existing minimum placement protection. The same GUI amount is used regardless of contract quantity—do not multiply it by the number of contracts.
Expanded Early Windfall Protection: Build1006 distinguishes between two separate Early Windfall systems. The Early Windfall Temporary Profit Target pursues a faster initial profit target during a configured bar window, while the Early Windfall Profit-Lock Stop attempts to protect a selected percentage of an early favorable move after the required peak-profit threshold is reached.
Persistent Early Windfall Profit-Lock: Once an Early Windfall Profit-Lock Stop is confirmed, expiration of its eligibility period does not cancel or loosen the protection. It remains active until filled, the position closes, or an equally protective or tighter stop is confirmed.
Improved Near-Exit Profit Lock: Evaluates the initial entry and every scale-in leg independently as the configured session exit approaches. Each leg must previously have reached—and must still retain—the selected amount of unrealized profit before its exact Near-Exit protective stop becomes eligible.
Improved PinnacleQuantRangeTracker™ AI Training Selection: Dynamic Mode now evaluates completed regular-session and overnight-session ranges, recent RTH expansion or contraction, overnight-range behavior, and the direction of the recent range trend when calculating AI Training Desired Profit.
Automatic AI Training Risk Selection: When PinnacleQuantRangeTracker™ Dynamic Mode is enabled, AI Training Risk Per Trade is calculated at 75% of the resolved AI Training Desired Profit, rounded up to the next whole dollar. Only completed sessions are used, keeping the calculated values consistent regardless of when the strategy is enabled.
Automatic HigherEdgeAI™ 3.0 and PQNews™ 2.0 Lifecycle: Both Python applications are launched, controlled, and shut down automatically through Price Action Pivoter™. Clients do not need to type commands into or manually operate either terminal.

Previous Version—V5AiBuild1005 (Released: April 2026)

Dedicated Python Application Management: Price Action Pivoter™ automatically launched and controlled the HigherEdgeAI™ 2.0 and original PQNews™ terminals for each configured instrument.

V5AiBuild1004 (Released: October 2025) – Now live for all Mach 15 and Mach 16 clients

Signal Caching for Faster Backtesting: Introduced the Speed Up Backtests by Caching Signals? option. Two-year HigherEdgeAI™ signal backtests that previously could take days could be completed in approximately 45 minutes, depending on the instrument, data, and computer.

Original PQNews™ Checker AI: Introduced the first Gemini-powered real-time news-awareness overlay using the client’s personal Google Gemini API key. At each configured polling interval, PQNews™ evaluated live web data, classified news impact as HIGH or LOW, and generated a news-confidence score for Price Action Pivoter™.

Configurable News Polling: Added the Poll News Interval (Minutes) setting, with a default interval of 15 minutes and a configurable range from 1 to 60 minutes.

News-Based Confirmation Support: Allowed qualifying high-impact news assessments to provide additional confirmation support when all other required Price Action Pivoter™ conditions were independently satisfied. PQNews™ did not originate trades or replace the strategy’s core entry logic.

Tactical Scaling: Added as many as three scale-in legs—L1, L2, and L3—beyond the initial entry. Separate operating modes supported scaling while the position was profitable or while it was experiencing a configured drawdown.

Configurable Scale-In Conditions: Added independent scaling thresholds, maximum thresholds, minimum bar spacing, optional price-action filters, protective-stop distances, and custom L1, L2, and L3 profit targets.

Rapid Rejection Guard: Added a tighter ATR-based or hard-dollar protective stop during the first configured number of bars after an initial entry. After the temporary window expired, the RRG stop could revert to the full MyRiskPerTrade distance unless another trailing-stop system had already established tighter protection.

Early Windfall Temporary Profit Target: Added an aggressive, time-limited profit target for initial entries using an ATR(5) multiplier, a hard-dollar target, or both. If the windfall target was not filled during the configured bar window, the strategy restored the normal MyProfitTarget order.

Near-Exit Profit Lock: Added protection intended to tighten qualifying stops as the configured session exit time approached, helping protect unrealized profit before the strategy’s scheduled flatten time.

Stop Competition—Most Protective Stop Wins: Introduced coordinated comparison among simultaneously active protective-stop systems. The strategy selected the tightest valid stop candidate—the highest valid stop for LONG positions or the lowest valid stop for SHORT positions.

Refined PinnacleQuantRangeTracker™ Dynamic Mode: Improved the automatic selection of AI Training Desired Profit and AI Training Risk Per Trade using market-range analytics rather than requiring traders to select both values manually.

HigherEdgeAI™ Scheduled Launch: Allowed HigherEdgeAI™ 2.0 to be scheduled to begin its machine-learning process before the first configured trading window.

Enhanced Signal Generation: HigherEdgeAI™ 2.0 delivers significant improvements in LSTM model training and pattern recognition accuracy.
Refined Signal Analysis: V5AiBuild1004 introduces advanced signal filtering logic.
PinnacleQuantRangeTracker™ Integration: This proprietary indicator analyzes actual market movement over 2-hour periods using multi-day statistical data to recommend optimal AI Training Desired Profit and AI Training Risk Per Trade parameters. The indicator can also be loaded separately on any chart by right-clicking the chart and selecting Indicators > PinnacleQuantRangeTracker.
M-ATR Dynamic Stop System: Adds intelligent volatility-based stop protection that activates based on dual thresholds (ATR(5) levels and unrealized profit percentage), working in harmony with the Peak High-Low trailing stop system.
Session-Aware Risk Scaling: Optionally automatically reduces the MyRiskPerTrade parameter when cumulative session profits exceed user-defined thresholds, helping preserve gains during profitable trading sessions.
Windows 11 Users: On your trading machine, go to Settings > search for “Terminal Settings” at the top > next to “Terminal” select Windows Console Host. This ensures the Python terminal that HigherEdgeAI™ 2.0 runs on behaves optimally.

V5AiBuild1003 (Released: June 2025)
Now live for all active Mach 15 and Mach 16 clients.

Refines signal analysis logic to be more lenient than Build1002, allowing more flexibility in entry conditions.
Adds a new consistency check to ensure signal quality remains high despite broader entry conditions.
Continues to use the Early Prediction Box for real-time speed improvements (up to 58 seconds faster), with all trade entries and exits still executed on bar close.
Trade count may increase over Build1002, but signal quality is maintained through enhanced signal filtering.

Prior Version — V5AiBuild1002 (Released: May 2025)

Integrated longer timeframe signal analysis to help avoid sideways price action.
Improved real-time signal speed by up to 58 seconds while Bar[0] was still forming when the Early Prediction Box was enabled.
Expected fewer total trades compared to Build1001, but with higher trade quality due to extended machine learning analysis.

Older Version — V5AiBuild1001 (Released: April 24, 2025)

Introduced AI Risk Per Trade Training, allowing the HigherEdgeAI™ model to train on a lower stop price than live trading, improving robustness.

Original Version — Price Action Pivoter™ V5Ai (Released Q4 2024)

Introduced HigherEdgeAI™ real-time machine learning model and improved price action strategy.

Questions? → Schedule a Live Zoom Demo for Price Action Pivoter™ V5Ai (Build1006):

About The Author

This article was authored by Raffi Sosikian, a highly proficient software engineer and enterprise architect. He has an MBA, holds a Series 3 license, and is the principal at Pinnacle Quant, LLC, CTA, a boutique commodity trading advisory firm specializing in building both custom trading systems (including private label for small funds to brand as their own software) and in-house pre-built quant and price action-based automated trading systems. As time permits, Raffi posts his financial analysis opinions on SeekingAlpha under the pseudonym “Prudent Research.”

I hope this powerful and unique automated trading system exceeds your day trading needs and serves you well for a lifetime!

Happy Trading, and God bless!

Free NinjaTrader Performance Dashboard You can Use With the Price Action Pivoter™ Software Suite

My favorite futures trading platform is NinjaTrader, and one of the small details worth understanding if you’re reviewing scale-in strategy results is how their Strategy Performance reports pair entries to exits. NinjaTrader uses FIFO (first-in, first-out) pairing by default, meaning that when a strategy scales into a position and later exits those legs, the report pairs them in chronological order — the first exit fill is paired against the first entry, the second exit against the second entry, and so on. The aggregate numbers (net profit, drawdown, equity curve) are 100% accurate; what can differ is the per-trade profit and loss attribution, because FIFO pairing doesn’t know which protective stop or profit target was originally attached to which entry leg.

I built the Pinnacle Quant Executions Dashboard to provide a complementary view for exactly this situation. It pairs each exit to its originating entry by order name, so a protective stop on the initial entry is shown against the initial entry, and each scale-in’s stop is shown against its own leg. Same totals as NinjaTrader’s report — but with per-leg attribution that matches how the strategy actually managed each position.

You can read more about NinjaTrader’s FIFO pairing convention directly in their official support threads here, here, and here.

The Pinnacle Quant Executions Dashboard works with any NinjaTrader strategy — simply export your Executions tab CSV and load it in. For Price Action Pivoter™ V5Ai, you get full per-leg attribution that matches exactly how each position was managed while trading. For other non-Pinnacle Quant strategies, the dashboard reconciles to NinjaTrader’s default FIFO pairing while still giving you a clean, professional-grade view of your performance. Either way, it turns your raw executions data into an informative performance dashboard instantly — no manual spreadsheets, no reconstruction, just an accurate picture of how you’re actually trading.

Your Privacy Stays Intact

I know how sensitive trade data is. This is a local HTML tool, which means when you drop your CSV in, the processing happens entirely on your own computer. Your data is never uploaded to my server or shared with anyone – it stays completely private and secure, as it is only running on your computer/VPS.

How to Use NinjaTrader Performance Dashboard

Download: Unzip the file, save the .html tool to your computer, double click on so it opens in your browser.
Export: In NinjaTrader 8 > Control Center > Executions > Right-Click > Export (as .CSV).
Analyze: Drag your CSV file onto the page, and you’re done! Enjoy a nice, clean, intuitive dashboard.

How The Dashboard Looks