FAQ

STOKAI is a Financial asset pricing software, designed with the goal to make sophisticated predictive algorithms, used at Investment Banks and Hedge Funds, available to individuals. Key applications of this tool is in trading stock index, FX/Forex and commodities CFD’s and other products like options, futures and ETF’s.

Backtesting is the general method for seeing how well a prediction algorithm would have done ex-post. It assesses the viability of the algorithm by discovering how it would play out using historical data.

Backtesting page for each asset has 11 graphs. Showing, each day’s prediction for last 11 days. The blue lines show what was predicted on that day. The x-axis shows the dates the prediction is done for in the future. And the red lines show the actual, observed market prices everyday.

For instance, the first graph on the backtesting page is for prediction done yesterday, so it will have only one red dot for yesterday’s actual market price. The second graph, is from day before yesterday, it has a red line with 2 points – observed market values yesterday and day before. The final graph has the red line with 11 points as we have observed market prices for each of the last 11 days.

Why 11?

Because each day’s prediction is done for today and 10 business days in the future. After today, today’s prediction graph moves to backtesting page and stays there for next 11 days. It moves down everyday as we progress through time and new graphs get added above. The Blue line (predicted value as of that day) will never change. It will be there as it was on the day of the prediction, to ensure a fair and honest view of how the prediction has performed.

In summary, we show the actual prices observed in the market(red lines) over the last 11 days plotted against the prediction on each day over last 11 days(blue lines). This allows the user to see performance of STOKAI algorithms.

Simulation or modelling is the action of building an abstract representation (a model) of a real world financial state(macroeconomic and/or microeconomic).

Simulation of underlying market variables(like bond yields, inflation rate, unemployment rate etc.) is done to produce the prediction/forecast of the financial asset(stock index, fx or commodities).

Future simulation is used for calculating projected P&L and for Risk Management of existing portfolio in the industry.

In financial services industry, simulated future prices are used to hedge existing portfolio and for pricing ‘Over The Counter’ derivative securities like bespoke options and forward contracts. 

Forward contracts are similar to Futures in their purpose, i.e. to fix price of something on a future date. Key difference being, Futures are exchange traded and have standardised contract size. So, the price is determined based on supply and demand on the exchange.

On the other hand, Forward contracts are bespoke products, so quantity and price is agreed between two parties using their own algorithms and legal contract details. Hence the term – Over the Counter.

The algorithms(like STOKAI) used to calculate this fair price, can also be used by Futures traders to speculate on the market. i.e. based on Forward prices, you can make an assessment of whether the Futures market is underpriced or overpriced.

Long term investors/pension funds rebalance their portfolios less often than day traders. for eg. quarterly, yearly etc. Therefore, this is a risk management tool for you, in addition to existing good judgment and mean-variance based statistical tools used in the industry today. Thus, you can use this tool to decide to exit or expand your position outside of asset re-allocation phases, if the algorithm output shows a significant movement in the near future . You can also use this to decide whether or/and when to reduce/increase exposure to a sector, country etc.

A macro-economic risk factor move that impacts the entire market index will most likely impact positions in single stocks as well. Based on beta correlation factor, this could be positive or negative move and by a smaller or a larger magnitude.

Indeed, as we grow, we will add more asset classes – Bonds and Rates, but not necessarily in that order.

Currently our focus is Equity composite indices, FX and Commodities, because they are the most popular markets.

We publish backtesting results of last 10 days, showing past predictions in blue lines against realised market price in red lines.

We publish predictions everyday Mon-Fri by 8:00 AM GMT for current day close price and onwards. So, for eg. prices published on Mon morning 8:00 AM GMT are for Mon close and onwards. 

No, it’s not. The simulated future prices are the output of the algorithm without any manual intervention. The output of the algorithm and the input data has no cleanup, adjustments or rounding applied.

No human bias is allowed in building the algorithms and hence the only gurus consulted are advanced calculus mathematicians when building the model.

The predicted future prices do not constitute financial advice in part or whole and does not represent view or opinion of the company or any of its directors, shareholders, customers or employees.

STOKAI algorithms are run using a snapshot of market data as of close of day yesterday. Hence, it cannot take current day’s market news into account. So, intra-day volatility after STOKAI algorithms have already produced the prediction output, is likely.

However, consistent deviation between predicted and realised prices every day, means the model believes that the market is either under or overpriced, and that it will correct itself over the long term. Buying/Selling Futures is the best way to profit from this scenario.

Finally, on most days while working with STOKAI algorithms, you will find that the end of day close price target is hit earlier in the day, in which case the close price may be different due to profit taking. Backtesting graphs, show end of day close price only and not high or low during the day and hence offer a very conservative & critical assessment of STOKAI performance.

We do not make tall claims about past performance. With any past success rate, we understand and want the user to also realise that future will still always remain uncertain and any position taken in the financial markets will carry risk, and hence the reward, if things go your way.

In addition, what is the correct way to measure success of past predictions? It is important to define a metric first. Do we consider the magnitude of losses and profit? Do we take into account the early close of the position? What if the profits were higher than predicted and losses lower than predicted, is this measure still a reliable metric? All these ambiguities can lead to a miscalculated success rate that will not mean much anyway.

Stokai is the output of a mathematical algorithm. This is not a signal or an advice to buy or sell a particular financial asset. Hence, when the markets are functioning efficiently based on fundamental economic variables, Stokai is good for guiding user on how to take short term position in the market for a quick profit. And when the markets are not working efficiently, STOKAI shows whether they are over or under priced.

Stokai is also good at showing the user, when the markets are extremely volatile and completely unpredictable, and hence, better left alone. Making sure trading in an unpredictable, volatile market is avoided, (thereby averting losses) is as much a success as profiting from a trade. However, it can be difficult to include such a metric in one number such as success rate.

Stokai works differently to other tools and it will take some time to grasp how to make the best use of it.

For FX and commodities, yes, because for eg. FX SPOT markets are operational 24 hours a day on weekdays. Therefore quoted prices seen on broker website reflects the actual prices in the global FX market. And since STOKAI is based on fundamental SPOT data, STOKAI forecast/predicted end of day close means the same as close of day prices seen on your broker’s trading platform.

To put it simply, ‘close of day’ cutoff has the same meaning for actual FX SPOT market, broker’s quoted prices and STOKAI’s predicted/forecasted prices.

For Stock Indices, STOKAI shows the predicted/forecast prices, as at closing time of the stock exchange, which is generally in the range of 4PM – 5PM local time for most global stock exchanges. However, in the current high speed internet age, and 24 hour retail trading environment, brokers quote bid and ask prices for stock indices, outside of official stock exchange trading hours.

In fact, ‘out of hours’ market trading occurs, for indices of almost all major stock exchanges around the world, and is not limited to institutional investors anymore. A few things are important to understand about this market.

• The prices quoted outside of exchange trading hours, on broker sites, are speculative/expectation of next day’s prices. This is because an index value is the weighted sum of mkt capitalisation of it’s constituent companies. And prices of constituent companies’ stock is not publicly available until the next day, when the stock exchange opens.
• The bid-ask spreads quoted by brokers, generally, tends to widen during out of hours trading of stock indices. This is done to account for the additional risk taken on by the broker, of adverse movement at ‘market open’, the next day.
• The ‘out of hours’ prices, quoted on broker sites, are not pulled out of the hat. They depend on a few factors, such as market news about constituent companies, for instance new deals, bankruptcy etc.; exchange traded futures and ETF’s prices at other exchanges that are open; and prices of other indices for eg. FTSE 100 out of hours price will take cue from Dow Jones and S&P 500, post close, and NIKKEI 225, pre-open.

These are the most liquid stock indices and FX pairs traded globally and hence the markets are efficient and more transparent. Therefore, the algorithms based on macroeconomic fundamentals can be built with a high degree of confidence and accuracy.

Nevertheless, if you are interested in algos for an index, commodity or FX pair not listed, please let us know, and we will add them to our development pipeline, for next release. Provided, we can come up with a reliable algo for that asset.

No, as that would not be fair to current members. Moreover, we have a grandfather/grandmother policy to our membership. This means, you will only ever be charged what you signed up for, and never be affected by future price increase. Unless you cancel and rejoin, when the prices have increased.

There will never be a cheaper alternative offered. Only exception being, annual/semi-annual/quarterly membership, where members pay upfront, and therefore may be cheaper than recurring monthly membership. Or other one-off promotions and marketing campaigns.

Indeed, backtesting of financial engineering algorithms, is done on data that is 5-20 years old. And that is exactly what we do, during the design, build and testing of STOKAI algorithms. However, for “day to day” trader’s or investor’s, trading decision making point of view, last 10 days backtesting results reporting is more relevant due to practicality and regime changes.

Firstly, to elaborate practicality, STOKAI algo shows prediction/forecast for 10 days in the future. Hence, it is more informative for the user to see how the last 10 days predictions, have played out in the markets. This helps user decide, whether to trade in that market short term/long term, or not at all.

Secondly, regime changes means, for example, that it’s more relevant for the user, trading Dow Jones during COVID-19 pandemic lockdown, to see the performance of STOKAI algo when Donald Trump is the president and US govt. is dealing with the pandemic. Rather than, when George Bush was the president, and US armies landed in Iraq and Afghanistan. The macroeconomic environment is very different, and govt. and monetary policies are also, completely different.

Regardless, as mentioned above, we backtest the algorithms during development, and buildout, with old data from past recessions, regime changes and stressed(high rates/low rates/high inflation/low inflation etc.) economic environments, to ensure that our algorithm will stand anything, the uncertain future may throw at it.

Seasoned traders, with a few years experience under their belt, have learnt(most of the time, the hard way) that markets do not present profitable trading opportunities, all the time and every day. Trading for the sake of keeping a position open, is not the right way. Also a signal or alert is not very informative about whether to take a short term profit position, or a long term position waiting for market correction.

Stokai algorithms provide fair value of the market, based on macroeconomic data, news and events.

The user then, has to make assessment of the direction to trade, and the length of time the position is to be kept open, based on the algorithm’s performance over the last few days. We always recommend using paper money trading accounts to new market entrants, while they learn the ropes.

Here is how, our long term customers do it-

1. Login to Stokai and go to Backtesting section. Here, you will see performance over the last 10 days. The blue line is what Stokai predicted and red line is what actually happened.
2. Then go through the markets and see where these two lines are close, and where, they are not.
3. When these two lines are close, that means market is working efficiently and when they are not close, that means the market is inefficient(i.e. either underpriced or overpriced).
4. Now you have two options
   • Use the graphs where the market is working efficiently(blue and red lines are close to each other), and take a short term position for quick profit. For eg. Close it before end of day or within the next day or two.
   • Or use the graphs that are not tracking the market closely(blue and red lines are off from each other) and take a long term position. For eg. 5 days to a month, waiting for correction.