a Polish FinTech startup


8888888b.                                                  888                                                       .d8888b.                                       
888   Y88b                                                 888                                                      d88P  Y88b                                      
888    888                                                 888                                                      Y88b.                                           
888   d88P 888d888 8888b.   .d88b.  88888b.d88b.   8888b.  888888 888  888  .d8888b 88888888 88888b.   8888b.        "Y888b.    .d88b.  888  888  888  8888b.       
8888888P"  888P"      "88b d88P"88b 888 "888 "88b     "88b 888    888  888 d88P"       d88P  888 "88b     "88b          "Y88b. d88""88b 888  888  888     "88b      
888        888    .d888888 888  888 888  888  888 .d888888 888    888  888 888        d88P   888  888 .d888888            "888 888  888 888  888  888 .d888888      
888        888    888  888 Y88b 888 888  888  888 888  888 Y88b.  Y88b 888 Y88b.     d88P    888  888 888  888      Y88b  d88P Y88..88P Y88b 888 d88P 888  888      
888        888    "Y888888  "Y88888 888  888  888 "Y888888  "Y888  "Y88888  "Y8888P 88888888 888  888 "Y888888       "Y8888P"   "Y88P"   "Y8888888P"  "Y888888      
                                888                                    888                                                                                          
                           Y8b d88P                               Y8b d88P                                                                                          
                            "Y88P"                                 "Y88P"

validating financial models with machine learning || developing collaborative intelligence trading systems

Kim jest Pragmatyczna Sowa?



Left Align

Pragmatyczna Sowa to polski startup FinTech, który realizuje projekty związane z wykorzystaniem sztucznej inteligencji w obszarach takich jak: financial backtesting, data analysis z naciskiem na anomaly detection oraz budowę systemów tradingowych opartych na collaborative intelligence.

Projekty realizowane przez startup rozwijane są z ekspertami z dziedziny rynków finansowych oraz technologii informatycznych. Przestrzeń projektów Pragmatycznej Sowy umiejscowiona jest wyłącznie w polskiej bezpiecznej chmurze.

Pragmatyczna Sowa pomaga firmom transformować się cyfrowo poprzez konsultacje i wdrożenia, aby technologia była katalizatorem dla rozwoju a nie blokerem. Kładziemy duży nacisk na cyberbezpieczeństwo swoje i naszych klientów. Pomagamy polskiemu biznesowi zwalczać cyberzagrożenia, chronić ich zasoby oraz dbać o cyfrowy wizerunek.

Our projects

DiNapoli AI

Project DiNapoli AI launched in late 2019 with a mission to employ the potential of artificial intelligence to support trading based on DiNapoli levels. Project is entirely developed in Poland by technology specialists and polish DiNapoli experts. The outcome of this project such as: tools and new solutions is addressed to the global DiNapoli traders community. The core of the project are two disruptive technologies: cloud computing and artificial intelligence. Since 2020, we have managed to reach collaborative intelligence status, where traders work together with AI in one team. Project received no external funding. Made by traders for traders.

The Project Team

  • the Polish Team of DiNapoli Experts: Piotr Grela, Pawel Besler, Tomasz Augustynowicz.
  • Michal Tys, IT engineer | a FinTech startup PragmatycznaSowa.p
  • Ceti | The Data Center team | www.ceti.pl

Just have a look how AI conducts its own analysis within a cloud environment. Scanning the Nasdaq with 3727 tickers took less than 4 min. As a result, AI has preliminary recognized 47 potential RRT patterns for further analysis by human traders.



Project website | DiNapoli.ai

FinancialMarkets.cloud
a cloud-based financial information system powered by Artificial Intelligence

***proof of concept / a working prototype (Spring 2020)***
***Alpha release coming soon***

Opis elementów składowych prototypu w języku polskim



Prototype at a glance:
  • it reads the market not the financial news!
  • gathers financial data from independent sources (global stocks exchanges, CFTC, FED, volatility indices, FRED, Eurostat)
  • performs analytics and data visualization
  • shows the current state of global markets and economy (overview of indices and macroeconomic indicators, industry sectors and stock portfolios)
  • uses AI to find strong predictive price patterns (DiNapoli's directional indicators) in stocks, indices and commodities, that might be early warnings of possible trend reversals
  • it spots anomalies (volume spikes, COT report, stock market sessions) and similarities
  • powered by Artificial intelligence
  • cloud based
  • accessible from everywhere via any web browser
  • analyses markets globally (Asia, Europe, US)
  • fully developed and maintained in Poland (GDPR friendly, Cloud Act free)
  • no software installation or data subscription needed
  • Artificial Intelligence trained by real traders

'the Railroad Man' (Kolejarz)
DiNapoli's directional indicator scanner powered by Artificial Intelligence

(Proof of concept / a working prototype, Spring 2020)


What is the Railroad Track directional pattern (indicator)?

According to Dinapoli's "Trading with DiNapoli Levels" book, the RRT directional pattern is one of the best directional signals that can be found on charts in any time interval. Typically, is accompanied by a high market volatility and depicts a rapid change in the price of a particular financial instrument followed by a relatively quick return to the point from where the rapid slope has started. For experienced DiNapoli traders, prompt identification of that pattern could be turned easily into profit.


Prototype at a glance:

  • it uses disruptive technology such as convolutional neural network instead of traditional programming or algorithmic trading
  • it is cloud based, hence it doesn't require your machine's computational power
  • the whole process is fully automated: data download, deep learning analysis and report preparation (PDF easily accessible via web browser). Process is triggered multiple times during a day (Mon -Fri) according to the time zones of the stock exchanges in Asia, Europe and the US
  • accessible from everywhere via any web browser
  • it analyses markets globally (Asia, Europe, US)
  • fully developed and maintained in Poland (GDPR friendly, Cloud Act free)
  • no software installation or data subscription needed
Project overview

In this project we are using deep learning, which allows us to teach the machine how to recognize the DiNapoli's directional patterns on charts. We are showing Artificial intelligence our RRT patterns (charts) that we have found manually beforehand, so AI can learn from those and find similar ones on the market.... As simple as that.

How the scanner works:

  1. OHLC data are downloaded to the cloud
  2. OHLC ticker's data are transformed into a bar chart
  3. Chart is transformed into a matrix (an array of numbers)
  4. Matrix enters to the pre-trained neural network for the analysis process
  5. Neural network tries to evaluate if the RRT pattern is present on the chart
  6. All charts with the high RRT probability go into a PDF report.

At this stage, the scanner has access to the following exchanges & markets:
  • AMEX,NYSE,NASDAQ,BATS
  • Cryptocurrencies, Global Indices, Commodities, FOREX
  • London Exchange, XETRA Exchange, Warsaw Stock Exchange, Borsa Italiana, SIX Swiss Exchange
  • National Stock Exchange of India, Shenzhen Exchange, Shanghai Exchange, Hong Kong Exchange, Thailand Exchange, Tokyo Stock Exchange


backtestBefore.trade | financial backtesting powered by machine learning




cot-report.online | a detailed analysis (inc. anomaly detection) of the weekly Commitments of Traders Report issued by CFTC.





Nasze artykuły

How to teach Artificial Intelligence to read the stock charts as humans do

FinancialMarkets.cloud | a cloud-based financial information system powered by Artificial Intelligence

How to teach Artificial intelligence to recognize highly predictive price patterns on the stock market.

Hunting hidden patterns in soft commodities with machine learning algorithms

Breaking down the NSL-KDD dataset and its predecessor KDD 1999

Employing Deep Learning for Cyber Security :: Artificial neural network (ANN) case-study

Employing basic CNN for image recognition (Deep Learning case study)





Positive SSL