Live trading Results
ServeQUANT Reference Account 


NOTE: The pip value calculated by MyFxBook for XAUUSD and USDNOK is 10x the pip value of other Forex pairs. So 100 pips on XAUUSD or NOKUSD displayed by MyFxBook would mean 10 pips. ServeQUANT HUD displays the the pip values correctly across all pairs.


The above chart is updated at 00:00 hrs GMT

The strategy has been trading GBPJPY, XAUUSD and USDNOK since December (during its pre-launch Phase 0). Additional pairs enabled since 10th February 2020 (Phase 1): NZDUSD, USDCAD, AUDUSD, EURUSD, EURJPY, GBPUSD, EURGBP, USDJPY, USDCHF, EURCHF and CHFJPY - this is in addition to GBPJPY, XAUUSD and USDNOK.

In Phase 1, the above mentioned 14 forex pairs will be monitored by the ServeQUANT Machine Learning Platform and Quant generation pipeline. Top three Quants providing the highest success probability score will be used to generate trades on a daily basis. This further enhances the strategies potential to produce high rewards at the lowest risk. The max drawdown level of the system will be limited to 500 pips.

ServeQUANT trading style can be observed using the price action charts below.
You can click on traded symbols to plot the trade entry/exit on the charts to observe the strategy's trading style.

Powered by Myfxbook.com

How it Works

Phibase team has been looking for solutions to deliver a high reward - low risk system. Clearly, the traditional algorithmic approaches have their limitations - Everything is based on the drawdown tolerance and the need to wait for suitable price action for the strategy to profit.

Machine learning has opened up new possibilities for our development team. Traditional software programs are built using rules and past data to obtain favourable results. Here we expect future price to follow the rules. In contrast, Machine learning enabled us to build systems which take in past data and results to create rules which could provide favourable results on future price. Markets are chaotic - they are truly random. serveQuant provides dynamic trading for the dynamic markets.

serveQUANT strategy creates price models for the next day, week and month using past data. Since the markets are random, prediction of future prices is impossible. The strategy is to generate multiple price action models using the machine learning platform. These models provide the system with several alternative price action paths that may occur in future. The price models are run through our automated logic creator. The logic & parameters with highest success probability across all predicted price models is selected and a Quant is created. The system is designed to scan over 30 Forex pairs for best possible trading opportunities. QUANT backtest over past 100 trading days using actual historical data. Quants with over 65% win-rate are delivered by serve QUANT to your MT4 in real time.

SQ-Pipeline
Click here to download presentation on serveQUANT Features
Click here to download presentation on presentation on  EAs vs Quants
Multiple Trading Pairs

ServeQUANT will explore trading opportunities across all major, minor and exotic pairs. The strategy has been trading GBPJPY, XAUUSD, USDNOK, NZDUSD, USDCAD, AUDUSD, EURUSD, EURJPY, GBPUSD, EURGBP, USDJPY, USDCHF, EURCHF and CHFJPY. Traded pairs will change dynamically. Minor pairs with higher spreads will only be traded if expected pay-off per trade is at least 10 pips.

Consistent Gains

The most important feature is that the system continuously adapts to the market conditions. You will not need to wait for certain price conditions or cycles for trading to be profitable. Making consistent gains has been a challenge - we believe serve quant will be able to deliver on this promise. The monthly gain curve shows potential for steady gains with a monthly win rate of over 90%.

Phibase Machine Learning

Phibase has been researching into machine learning for several years now – iProfit, our self contained neural network based strategy which has proved itself to be extremely robust. While it has been chugging along, the gain to loss ratio is less than satisfactory. However, this strategy has proved how neural networks can be used successfully.

SQ-ML

Artificial Intelligence is a general field which was popular in 1980s with approaches like Symbolic AI and Expert Systems. Such approaches were good for solving well defined, logical problems like Chess. These approaches were unsuitable for complex, fuzzy and real world cases like image or speech recognition.

Machine Learning is a new approach where the system is trained rather than being programmed with hard-coded rules. A ML system is provided with example data and expected results to come with a new set of rules for the task at hand.

drive01

Deep Learning is a specific sub-field of ML which puts emphasis on learning successive layers of increasingly meaningful representations of the input data. Deep Learning is a multistage way to learn data representation – This makes it most suitable for price prediction.

>

serveQUANT : Trading Strategy

Phibase development team has implemented the Machine Learning Platform which is a complete pipeline to predict future price moves based on machine learning (ML). Serve QUANT will explore trading opportunities across all major, minor and exotic pairs. Currently tests have been completed and Pre-launch Phase Zero starts trading with USD NOK, GOLD and GBP JPY. Pairs with sufficient price range and volatility at present. Traded pairs will change dynamically. Minor pairs with higher spreads will only be traded if expected pay-off per trade is at least 10 pips.

  • The most dynamic trading strategy
  • Logic and parameters are based on multiple predicted price path scenarios – Not actual past data
  • Logic tested over most recent 100 days or 100 trades on actual historical data for validation
  • Logic selected only of recent 100 trades delivered gains at-least twice the drawdown depth
  • Logic with over 300 pips DD in 100 day backtests are rejected
  • Quant created for selected logic and runs on serveQUANT master account
  • Quant triggers trades on master account and client accounts instantly.
  • 100% trade match across all brokers – Quant uses price feed from broker on which testing was done.

serveQUANT : Advantages

serveQUANT is dynamic – Not passive. Does not wait for suitable price action to be profitable. Logic is adapted to predicted unseen price action models and tested on recent actual price.

  • serveQUANT’s logic is adapted on a daily/weekly basis
  • No long duration drawdown periods or deep drawdowns
  • Max Drawdown expected is less than 450 pips
  • Trades between 20 to 30 times per month
  • serveQUANT triggers trades on both master and client account instantly
  • 100% broker independent – US broker compliant
  • NO GRID/NO MARTINGALE
  • High reward – Low risk system. Ideal for low leverage ESMA and US regulated accounts.