What is Ainslie Intelligence and what are
the Trading Signals? Ainslie is one of Australia’s leading gold
and silver bullion dealers, since 1974. Since mid-2017 we have been one of the largest crypto-currency
over-the-counter dealers. Ainslie Intelligence is simply an extension of our business, aimed
to provide better service to our highly valued customers. The Ainslie Intelligence Trading
Signals is the result of ongoing research and development into state-of-the-art, evolving
analysis of the crypto and bullion markets. The aim of our AI is to identify the optimum
time to buy and sell for maximum profit, whilst taking the lowest risk when modelled on average.
Originally, at the start of the year we used the AI as a part of our own internal trading
but have since transformed it into a product exclusively for Ainslie clients. Our complex,
deep learning AI recommendations are delivered via email at 2pm, 7 days a week. We also proudly
offer our signals in AUD to better suit our customers.
What differentiates your signals from other similar products like trading bots?
The Ainslie Intelligence AI utilises multiple proprietary data sources and sophisticated
deep learning neural networks to create a model that evolves over time as the markets
do. The first distinction between our AI products and others (such as “automated trading bots”)
that can be found online is that the technology behind our AI involves extremely sophisticated
deep learning neural networks that are continuously working (24 hours a day) searching for patterns
and connections on multiple layers, and across multiple markets. They are always tweaking
and improving models and trading strategies, which is necessary because the market is changing
and evolving all the time. What is the process used to construct the
trading signals? The AI systems, supported by their specialist
human technicians, apply the following advanced processes to develop precise buy and sell
recommendations: Firstly, Various proprietary data streams
(across different markets) are provided that may be related to the price movement of the
chart under investigation. Specific proprietary indicators are provided, including traditional
indicators, zero lag indicators, fuzzy logic indicators, adaptive indicators, clustering
indicators, neural indicators and prediction signal indicators.
From there, AI genetic algorithms and evolutionary optimisation strategies are used to fine-tune
the indicator variables over a very large number of iterations. Then, AI deep learning
neural networks are used to find patterns in the data streams (including multi-dimensional
patterns too intricate for a human to identify), with the help of the optimised indicators
contributing in optimised combinations, to produce the most profitable tradable models
with minimal trading risk. Each model is then extensively tested using
unseen data to check performance reliability, and abandoned or advanced for further improvement,
dependent on the result. The final selected model is used to create a prediction into
the future about what is most likely to happen next, and the appropriate recommendation is
made (buy, sell or hold). At that point, the previous learnings are
captured as a starting point and the evolution begins again when new data is added (daily).
A virtual “panel of experts” is formed, utilising several of the most recent AI predictions
to form a majority prediction that is then presented on the daily chart as either a continuation
or change (new buy or sell signal) recommendation. This is a highly complex and computationally
expensive procedure, that requires distributed processing across multiple cores of dedicated
servers, running 24 hours a day. How do I interpret the charts and what do
the buy and sell recommendations mean? A recommendation remains in place until a
new buy or sell is generated in the opposite direction. The AI is attempting to predict
the best way to make the highest percentage profit gain, while taking the lowest trading
risk, when modelled on average. This can sometimes be confusing, as it may
involve predicting that remaining in a longer term “Buy” or “Sell” position, when
the market is moving against you, is better than switching and trying to capture every
small price change. There is lots of logic backing every signal, but it may not always
be obvious to see on the chart. How are your performance figures calculated?
There are an unlimited number of ways you can measure the performance of the charts.
Since the recommendations are accurately recorded on the day they are generated, you can directly
assess their reliability yourself. We include a percentage performance figure with each
chart, which is the simple addition of idealised percent gains and losses made by each recommendation
over the past 12 months. The best way we have found to show performance against the market
is to compare our performance to the performance of the market if you were to buy and hold.
This is provided with the Trading Signals every day.
How old is the data used to produce the Trading Signals and does the age of the data impact
on the profitability? Our trading day ends at 10am (0:00UTC). We
receive our data from a variety of 3rd party sources between 10am and approximately 12pm.
We update all our modelling with new data, and produce, ready to distribute, the trading
signals between midday and 2pm. This means the new data in the trading signals is always
4 hours old when distributed at 2pm. We don’t believe this impacts profitability negative
on average, based on the time frames the AI is working with. That being said, it is labour
intensive to meet the 2pm deadline and we do it as fast as possible to distribute
How do I interpret the recommendations for hybrid pairs like the GSR?
The Gold-to-Silver ratio is simply the price of gold per ounce divided by the price of
silver per ounce. It can be used to show the relative strength between the two precious
metals. If there is a buy on the GSR, the AI is predicting gold to outperform silver
and conversely if there is a sell on the GSR, the AI is predicting silver to outperform
gold. This can be used as a tool to guide the balance
of gold and silver in your portfolio or the ratio can be directly traded on platforms
that allow it. Same goes for Ether Price in Bitcoin. If there
is a buy on the Ether priced in Bitcoin, the AI is predicting Ether to outperform Bitcoin
and if there is a sell, the AI is predicting Bitcoin to outperform Ether.
What should I do when recommendations turn out to be wrong?
Financial markets are driven by a myriad of complex variables that are frequently changing.
It is quite unreasonable to expect that any model will be 100% right 100% of the time,
or anything even approaching that goal. We believe that success comes if you can build
some models that make money most of the time. The idea is to build a system that makes more
money on average than it loses. The biggest advantage that the AI behind our Trading Signals
offers is that it is not a static trading system. It is dynamic and always learning.
If it continued to be wrong over the long term then it would be either a truly unpredictable
market (which we consider to be unlikely as all markets are simply a reflection of their
participants and we believe market participants can be modelled) or we are not giving the
AI the information it needs to find tradeable relationships. We continue to work tirelessly
to improve and refine the data and systems, from multiple and diverse sources, over time
to solve this problem. The AI cannot actually see into the future. It is making the best
“educated prediction” it can make, based on all of the information it is provided.
We believe the Trading Signals are just one tool necessary for surviving difficult markets
and remaining in the game long enough to capture the big profit runs when they arrive.
Is there any way to have additional markets I am interested in added to the Trading Signals?
We are limited by the number of markets we can effectively analyse and report on based
on time and resource constraints. We have selected the markets we needed the analysis
for ourselves, that also best suit the needs of our predominantly Australian clients (although
international clients are also invited to contact us for further information). As we
acquire a larger team and even more powerful AI hardware into the future we may consider
adding more. If you would like us to consider a particular market you have in mind, you
are always welcome to get in contact and we will take your feedback into consideration.