Among the quantitative private equity managers who started locally, those who are good at T0 and high-frequency strategies must have the name of this institution. Wenbo Investment was established in 2014 and started with high-frequency futures. The current management scale is 8.5 billion. Since March this year, Wenbo’s excess has been repaired very quickly, and its 1000 index increase has jumped to the top three. Fengyunjun was fortunate to invite the strategy director of Wenbo recently, and had an opportunity to exchange and dialogue.
Q: In 2022, Wenbo’s 1000 indexed products will be ranked at the forefront of the quantitative first echelon. I heard that this year has made a large strategy iteration. Can you talk about how it is done?
A: Our index growth strategy was first to copy the index + T0 strategy. Before September 2021, the fundamentals will account for half and half of the volume and price; the market will be closed after the fundraising in July and August. After an iteration in the fourth quarter of last year, the volume and price accounted for 70%, and the fundamentals accounted for 30%. In February and March of this year, we carried out another iteration, adding strategy logic, and adding the logic of acquiring institutions to adjust positions at the bottom layer. There is now increased effectiveness and enhanced capture of agency behavior. At the same time, the logic of factor crowding degree control is also added. There are also more iterations of the continuous version recently, and we are more confident in the volume and price at present. Recently, the signal of minute frequency is being done, and the excess capacity will continue to increase.
Q: As for Wenbo, the public investors may be more familiar with our T0 and high-frequency areas. Then, what do you think is the core advantage of Wenbo compared to other quantitative funds in the market? For the current leading institutions What about the crowded volume-price factor?
A: Wenbo belongs to the local practical school, a quantitative private equity manager who has experienced the growth of the domestic A-share market. We are relatively strong in time series forecasting and timing; we formed an alpha data team in 2019 to form a characteristic strategy of medium and long-term fundamentals Alpha + short-cycle T0, and continuously expand and improve factors and combination models to form A methodology with stable Bo characteristics.
In addition, we started with high-frequency futures and accumulated many years of advantages in high-frequency and T0 . The accumulation in the high-frequency field has a strong inspiration for our follow-up work, such as the application of high-frequency data and high-frequency thinking in factor development. Like we have done a lot of short-cycle volume-price factors and high-frequency aggregation factors this year to capture institutional behavior. Our factor mining method is not the same as other companies. We may start from the order department and high-frequency data to reduce the frequency of these factors; for example, we have synthesized a lot of capture institutions, hot money, retail investors through tick data. Behavioral factors, weekly earnings tracking of these factors, and congestion detection on them. In addition, we will dynamically adjust the weights of factors according to market conditions, and will also do some timing for low- and medium-frequency factors. From this perspective, our factors have high heterogeneity and correspondingly low crowding.
Q: Can you tell us about the current strategy line situation? We know that after September 2021, there will be obvious style switching in the market and a decline in market trading volume. How will Wenbo deal with this situation?
A: At present, the strategy line mainly refers to the increase and hedging strategy. It means that we have 300, 500, 1000, Kechuang and Shuangchuang. For hedging, it is mainly a stable series of neutral products. The quantitative industry started in September last year, and all major quantitative managers have achieved different degrees of excess drawdown. From September to October, the managers with high price-volume ratio and price have a large drawback; from the end of October to December, the main Strategies that are biased towards fundamentals and other medium-to-long-term cycles will be at a disadvantage. In the proportion of models last year, analysts unanimously expected that the proportion of factors, fundamental factors and momentum factors in the medium and long-term cycle will be significantly higher than that of peers, resulting in the increase of the Wenbo 1000 index. The excess drawdown in the fourth quarter of last year mainly occurred. in the mid-October to December period. At present, the strategy has reduced the weight of partial basic factors, and the current strategy has a relatively high proportion of price and volume factors.
Q: Please introduce the staffing of the company and core investment and research personnel. How often will iterations and upgrades be done? As the management scale grows, how much does the company invest in hardware and people each year?
A: Our current research team plus IT and operation and maintenance team has more than 60 people, accounting for more than 70% of the total number; 90% of the investment and research are from the top 5 domestic universities. A number of gold and silver medalists in national and world Olympiads in mathematics, physics, chemistry, ACM, etc.
Specialist in Machine Learning was previously a Distinguished Fellow at IBM. The IT team is mainly from senior engineers from the Internet at home and abroad.
There will be strategy iterations and research progress reports every quarter, and there will be continuous strategy upgrades and optimizations every week and month. The company spends about tens of millions in hardware and personnel every year.
Q: How is risk control management done? How to ensure the validity and stability of the model when the market style is frequently switched?
First of all, the model is constantly iterating , and secondly, we have a relatively large factor library, which we are tracking all the time. In addition to the factors currently used in the real market, we will also keep track of some other strategies that are not listed in the real market, and they are also waiting to be used by us at any time. By constantly iterating on new factors, we can keep our model’s inefficiency behind, which is the level of the model.
In terms of final combination, we also have relatively high requirements for risk control . In this way, we can maintain relative stability when the market style is rapidly switched. For example, for some products with more stable targets, such as the stable series products, we control the excess volatility of the product very strictly, and have very high requirements for its drawdown.
Q: Please introduce the capacity and scale of the current main strategy lines, what is the future development plan, and will the scale be strictly limited? In addition, investors will pay more attention to the impact of the increase in scale on the excess. What do you think of this?
A: At present, the product line mainly refers to the growth and neutral strategy, and the strategic reserve is aimed at undertaking a scale of tens of billions. On the one hand, when we iterate on the strategy, we will stress test the entire strategy. We will test a larger scale than we expected to see if our strategy will deteriorate at a larger scale, including the model. What is the decay rate of , we will make some approximate calculations. On the other hand, our R&D team is also constantly iterating on the model , which also means that we will keep increasing the dimension and heterogeneity of the strategy, which will lead to a larger capacity of the strategy.
In addition, we are also developing many active strategies , or large order strategies, which may have a positive effect on the trading capacity. For example, for reversal strategies, the more orders you place, the lower the cost, and the impact on revenue is actually a positive contribution. But some, such as trend-based strategies or factors, the more you buy, the higher your slippage cost and price. When the scale reaches a situation where we think there will be excess attenuation, we will stop fundraising until we iterate on a more complete strategy.
Q: What is the current scale and capacity of Wenbo 1000 Index? How crowded is the current 1000 index increase?
A: The 1000 index increase is currently close to 3 billion, and the capacity is relatively sufficient. The scale of continuous iteration of the strategy will also expand accordingly. At present, according to the data calculated by the agency, the current product excess correlation of 1000 refers to the lowest among the three strategies of 500 and 300. I remember it is about 0.2, and the current crowding degree is not high.
Q: Many investors worry that under the trend of institutionalization of retail investors in the future, there will be a continuous decline in excess. What do you think of this problem?
A: When there are fewer and fewer retail investors and more and more institutions, it shows that the difficulty of obtaining excess returns is increasing. At this time, the competition is to quantify the innovation ability and continuous iteration ability of the investment and research team of institutions. In fact, we will find that every few years, the style of the market will undergo a round of relatively large changes. When a team cannot continuously track the performance of the market, or cannot have a continuous innovative methodology, the company’s performance It will drop sharply and be gradually eliminated by the market. So, we are very important in this one.
Therefore, our company pays special attention to the staffing and some cutting-edge trading concepts and methods . Now every week, we will discuss cutting-edge trading ideas and methods, and develop these trading ideas into strategies for storage. In this way, in the face of crowded traditional factors, we will have many alternatives, and through these innovative researches, we can improve the competitiveness of our strategies. In addition, we have done better and more timely in strategy iteration. For example, the fundamental strategy examines the excess attenuation on a monthly basis, and the volume-price model measures the effectiveness of excess excess on a weekly basis.
In the communication with Wenbo, Feng Yunjun summarized several characteristics of Wenbo:
First of all, from the perspective of the investment and research team, the talent pool is ahead of schedule and has a forward-looking vision. The two founders have always been in the front line of investment and research, and they attach great importance to innovative research methods and cutting-edge ideas, which also makes Wenbo rich in strategic reserves and constantly enhances its strategic competitiveness;
Second, be good at amplifying your advantages and continue to iterate. One of Wenbo’s advantages is reflected in the accumulation of T0 and short-cycle strategies for many years. It has its own unique advantages in the development of volume and price factors, and has strong strategy iteration capabilities. For example, more heterogeneous factors can be developed, especially on the more sensitive trading volume such as the 1000 index increase strategy line, the effect of high frequency enhancement is more significant.
Finally, pay attention to risk control management and have strict control over excess volatility.
Thinking, innovation, forward-looking strategy development and iteration, and strict risk control are the most important points for a quantitative manager to survive and develop and have core competitiveness.
$ Wenbo 1000 Index Enhancement No. 1 B (P001223)$ $ Wenbo Zhongrui No. 6 (P000995)$ $ Wenbo Houxue Long Slope Quantitative Neutral No. 1-3 (P000988)$
This topic has 1 discussion in Snowball, click to view.
Snowball is an investor’s social network, and smart investors are here.
Click to download Snowball mobile client http://xueqiu.com/xz ]]>
This article is reproduced from: http://xueqiu.com/9047540546/228131119
This site is for inclusion only, and the copyright belongs to the original author.