We are a firm connected with the Forex Industry. We have millions of trading data like instrument, demographic, session, challenge phase trader’s trading, funded traders trading, passing rates etc. For example, the most traded pairs are XAUUSD, EURUSD, GBPUSD, US30. The most profitable traders are from Vietnam, UK. The most profitable trading session is London. Also we have trader wise metrics, like what is the good range of Win vs loss ratio, Max Losing Trade vs Average Winning Trade etc.
Let’s say traders from Indonesia takes payout 80% of the time when they get funded. We have these kind of data on many criteria.
Phase 1 traders care less about their trading and losses in most of their accounts, and when they pass into Phase 2, they get a bit serious, and their trading style changes with the excitement of getting funded. When they get funded, their mentality changes again as they want to secure some payout or at least recover the subscription fee. After they get their first payout, some traders trade on luck or risk aggressively as the fund is not their own. We need to figure out these analogies in more depth.
Secondly, let’s break it down into packages. Small account traders & big account traders trade differently. Another fact is that we need bigger funds to back up bigger traders & smaller funds for small account traders. So, if we break down challenge-wise, package-wise & also account-wise, the breakdown tree is scattered just here. Additionally, if we break it down by country, for example what is the data in Australia, Argentina & 100 other countries? Now this is very difficult to summarize. This happens for every package.
Then, let’s say we identified some traders based on these criteria. For example, we took the traders from Australia again. When we’re diving into their trading data, there are 50-60 symbols. Let’s say there are 10-15 most traded symbols that have their patterns/data like which are profitable or losing symbols among traders. Then, when the trader is taking trades, there are some real-time behavioral patterns. Like what lot size, symbol, and trading session is he choosing? If we even break down these 3 criteria, then again there are multiple more branches which is an issue.
Is it making a bit of sense? Now, is it possible to develop something out of it utilizing these data? Like the machine learning will use these data to understand what works and what not and then develop an algorithm. It will identify patterns. It will take trades based on the most profitable pair, session, metrics etc. Is it possible from your expertise?