Importance of Trading Data in Vietnam
In the Vietnamese derivatives market, API data, live boards from data providers are not actual order matching data.
The table below is an example data source that returns VN30F1M at 11:29:58, matching 820 contracts at a price of 1082. This data can be interpreted as around 11:29:58, there are 820 contracts reached at the 1082 range. The standard margin of error is usually around 0.3 points when the market has low fluctuation. However, it can go up to several points in the case of high volatility. For example, the deviation can be up to 30 - 40 points in the case of “call margin”.
The cause of this deviation is due to the phenomenon of “tick aggregation” - adding up many ticks in a period of time and returning a single tick. In Vietnam, the ticking time usually falls in about 02 seconds. The data returned can also be delayed by several minutes or even be unavailable for a period of time due to usually unpublished reasons. This means that, even with access to the best data source in Vietnam, the data still has deviations from reality. Therefore, traders may consider using algorithm-generated trading data to collect hidden data with absolute precision.
Conditions to Collect Trading Data
The necessary condition is a high-frequency algorithmic trading system. The data is returned continuously enough to reflect the general market, especially at sensitive times when the actual matching price and the hypothetical price highly deviate.
The sufficient condition is that the system should have many algorithms enough. Algorithm diversity will picture a market overview, thereby detecting favorable market conditions for each algorithm and vice versa. Furthermore, trading algorithms need to have a variety of strategies to avoid collecting the same data. Various trading strategies with different order types will enrich trade data. A typical example is the momentum strategy will give market execution data while market making strategy will give limited order matching data.
Practical Application
The first application of order matching data is to estimate the extent of market slippage. From thousands of orders matched, slippage can be calculated with high confidence. It helps increase the reliability of backtesting and paper trading. It should be noted high slippage is detrimental for algorithms using market orders but advantageous for algorithms with limit orders.
Traders should also get statistics of “sensitive” times of the market when the algorithms operate at a completely different speed compared to the average. At these times, the transaction frequency can be up to 10 times higher than normal. Any system improvement will bring outstanding results for automated trading.
Another important application of order-matching data is to form market expectations, thereby optimizing capital allocation to different trading strategies. An example is for the momentum trading strategy and mean-reversion strategy.
The last is to catch unusual transactions in the market, bringing operational advantages to the system. The principle is that if an unusual event occurs frequently, it’s likely an important signal to follow.
Combine market data and trading data helps ALGOTRADE adapt to different scenarios. For example, our system still operated reliably when live boards and market data in Vietnam were not stable during the Covid-19 pandemic.