1. Report Industry Investment Rating - Not provided in the content 2. Core Viewpoints of the Report - The report introduces the LSTM time - series model to improve the prediction of Treasury bond futures inter - period spreads without expanding the original factor set, finding that introducing time - series modeling can improve return stability and risk control for some varieties, but the sensitivity to time - series information varies among different contracts [38][39][40] - The LSTM model outperforms the DNN model on T and TF contracts, while the DNN model shows stronger prediction ability on TS contracts; the performance of both models on the TL contract is limited due to the short listing time [38][39] - Using the LSTM model's prediction probability for position weighting can enhance the return - risk characteristics of strategies on T, TF, and TS contracts, but the effect is not obvious on the TL contract [38][39] 3. Summary by Relevant Catalogs 3.1 Inter - period Spread Influencing Factors and Indicator Selection - Traditional machine - learning models and DNNs have limitations in time modeling as they assume sample independence in the time dimension and rely on manual feature engineering to introduce time information, and DNNs lack explicit sequence structure [7] - LSTM is an improved form of RNN, introducing a memory unit to decouple long - term information storage and current state output, and using gate - control structures to control information flow, which can better capture long - term dependencies [7] - Treasury bond futures inter - period spreads show trend characteristics in some stages, which are difficult to capture by previous models. Therefore, the LSTM time - series model is introduced to enhance the description of the time - series structure [12] 3.2 Recurrent Neural Network Testing Process 3.2.1 Data Processing and Sample Construction Process - The data includes fundamental factors of T, TF, TS, and TL contracts, and the features are constructed by aligning and introducing capital - related indicators and creating derived spread variables. The label is the first - order difference of the inter - period spread [16][17] - Data pre - processing involves removing early - listing samples and some end - of - month trading days, filtering small - amplitude spread changes, filling missing values, and standardizing data [18] - Time - series samples are constructed using a sliding window with a 5 - trading - day historical window, and samples are divided into training, validation, and test sets by strict time segmentation [19][20] - The model is trained using a weighted cross - entropy loss function and the Adam optimizer, with learning - rate decay and early - stopping mechanisms based on validation - set loss [22] 3.2.2 Parameter Setting and LSTM Network Structure - The task is a binary - classification prediction for the next - trading - day direction change of the inter - period spread, with input as a sequence of factor values over 5 consecutive trading days and output as binary logits [24] - The LSTM network has 3 layers, a hidden - state dimension of 8, and a Dropout ratio of 0.3. It uses the hidden state at the last time - step for classification [25] - The weighted cross - entropy loss is used to address class imbalance, and the Adam optimizer with learning - rate decay is applied for parameter updates [26][27] 3.3 Model Test Results 3.3.1 Comparison of Out - of - Sample Tests between LSTM and DNN Models - On T and TF contracts, the LSTM model has higher cumulative returns, Sharpe ratios, and better drawdown control compared to the DNN model; on the TS contract, the DNN model performs better; on the TL contract, the performance of both models is limited [28][30] 3.3.2 Probability - Weighted Backtesting of the LSTM Model - The prediction probability of the LSTM model is used as the position signal, and the position is normalized and limited to control risks [33] - Probability - weighted strategies improve the out - of - sample performance on most contracts, enhancing returns and Sharpe ratios without significantly increasing drawdowns, but the effect is not obvious on the TL contract [35] 3.4 Conclusion - Introducing the LSTM time - series model can improve the prediction of Treasury bond futures inter - period spreads for some varieties, but the effectiveness depends on contract characteristics, sample coverage, and spread structure [38][39][40] - Using prediction probability for position weighting has potential in improving strategy performance, and future research can further test the applicability of time - series modeling under more complex conditions [39][40]

1. Report Industry Investment Rating No relevant content provided. 2. Core Viewpoints of the Report - The stock index futures market is in an upward trend, with electronics and non - bank finance contributing to the rise of major indices. The basis of each variety has strengthened significantly, and trading volume has increased month - on - month. For bond futures, the IRR of bond futures has declined this week, and the inter - period spread has been oscillating strongly. The commodity market has seen the profitability of term structure and trend momentum factors weaken, while volatility, term basis, and warehouse receipt factors have performed well [3][55][77]. 3. Summary by Relevant Catalogs 3.1 Stock Index Futures Quantitative Strategy Tracking - **Market Review**: The market is on an upward trend. Electronics and non - bank finance contribute to the rise of CSI 300, SSE 50, and CSI 500 indices, while electronics and power equipment contribute to the rise of CSI 1000 index. The basis of each variety has strengthened significantly, and trading volume has increased month - on - month. IC and IM remain in a contango state [3]. - **Basis Strategy Recommendation**: Due to market sentiment, the basis of each variety has strengthened significantly. In the case of increased market volatility, the impact of market sentiment on the basis increases. For inter - period positive spreads, beware of the risk of large - scale fluctuations in the basis of far - month contracts caused by market speculation. The inter - period momentum signal recommends IC inter - period positive spreads, and the IM inter - period signal turns to reverse spreads. The roll - over strategy recommends holding near - month contracts to avoid short - term basis fluctuations caused by market conditions [3]. - **Arbitrage Strategy Tracking**: In the inter - period arbitrage strategy, the net value of the strategy last week showed mixed results. The annualized basis rate factor made a profit of 0.8%, while the positive spread and momentum factors lost 1.6% and 1.4% respectively (6 - times leverage). The annualized basis rate factor mostly gave reverse spread signals. The net value of the inter - variety arbitrage time - series synthetic strategy lost 0.5% last week, with losses mainly contributed by IF/IH and IC/IM pairings, and the IC/IF pairing made a profit. The latest inter - variety signal recommends a 100% position to go long on IC and short on IF, and a 50% position to go long on IM and short on IC [4]. - **Timing Strategy Tracking**: All models of the daily timing strategy lost last week. The single - factor equal - weight, OLS, and XGB models made a profit of 0.1%, lost 1.6%, and lost 0.8% respectively. The latest signal of the timing model shows that the bullish signal has strengthened. The XGB model is bullish on CSI 300 and CSI 500, and bearish on SSE 50 and CSI 1000. The OLS model is bullish on SSE 50, CSI 300, and CSI 500, and bearish on CSI 1000 [5]. 3.2 Treasury Bond Futures Quantitative Strategy - **This Week's Strategy Focus**: In terms of basis and inter - period spreads, the IRR of bond futures has declined this week, and the inter - period spread has been oscillating strongly. The subsequent positive spread space is limited, and the inter - period spread is expected to oscillate. The interest rate timing signal predicts an upward interest rate, and it is recommended to choose high - duration varieties for hedging. The multi - factor timing strategy signal is neutral. The inter - variety arbitrage strategy signals for TS - T and T - TL are both bullish. The credit bond neutral strategy currently holds the 1 - 3 - year index with reduced duration and hedges with treasury bond futures [55]. 3.3 Commodity CTA Factor and Tracking Strategy Performance - **Commodity Factor Performance**: Last week, the domestic commodity market generally continued the previous week's trend. The number of rising and falling futures products was basically half and half, and the overall risk preference slightly increased. The profitability of term structure and trend momentum factors continued to weaken and declined slightly last week. The best - performing factors were volatility, term basis, and warehouse receipt factors. In the short term, pay attention to the callback of CTA strategy returns caused by trend reversals [77]. - **Tracking Strategy Performance**: Different strategies have different performance indicators. For example, the CWFT strategy has an annualized return of 9.3%, a Sharpe ratio of 1.58, a Calmar ratio of 1.06, a maximum drawdown of - 8.81%, a return of 0.39% in the recent week, and a return of 1.44% since this year [78].

1. Report Industry Investment Rating No relevant content provided. 2. Core Viewpoints of the Report - The market showed an upward trend last week, with different sectors contributing to the gains of various stock indices. The trading volume of each futures variety decreased month - on - month, and the basis weakened. IC and IM maintained a deep discount state. The report continues to recommend a positive arbitrage direction for cross - period arbitrage and roll - over operations [3][4]. - The performance of commodity factors was mixed last week. The price - volume trend factors declined slightly, the term structure factors rose slightly, and the basis and warehouse receipt factors fell slightly. The report is still optimistic about the performance of commodity CTA this year [80]. - For Treasury bond futures, the basis fluctuated narrowly, and the cross - period spread rebounded slightly. The capital interest rate continued to decline. The report suggests paying attention to the positive arbitrage and cross - period positive arbitrage strategies of Treasury bond futures [60]. 3. Summary by Relevant Catalogs 3.1 Stock Index Futures 3.1.1 Market Review - The market rose, with electronics and non - bank finance contributing to the rise of the CSI 300 Index, banks and electronics to the SSE 50 Index, electronics and non - ferrous metals to the CSI 500 Index, and electronics and communications to the CSI 1000 Index [3]. - The trading volume of each variety decreased month - on - month, and the basis weakened. IC and IM remained deeply discounted [4]. 3.1.2 Basis Strategy Recommendation - The basis fluctuated, and IC and IM maintained a deep discount. The current basis environment is driven by neutral short - hedging demand. The report recommends a right - side approach for cross - period arbitrage and roll - over, maintaining a long - near and short - far positive arbitrage direction [4]. 3.1.3 Arbitrage Strategy Tracking - In cross - period arbitrage, the net value of each strategy was flat last week. The annualized basis rate, positive arbitrage, and momentum strategies had profits of 0.3%, 0.5%, and 0.5% respectively [5]. - The signal of the cross - variety arbitrage timing strategy turned to long small - cap and short large - cap. The synthetic strategy had a profit of 0.2% last week. The latest signals suggest a 50% position for long IC and short IF in the IC/IF strategy and a 100% position for long IM and short IC in the IM/IC strategy. The cross - variety arbitrage cross - section strategy had a loss of 0.14% last week [6]. 3.1.4 Timing Strategy Tracking - The performance of the daily timing strategy models was differentiated last week. The single - factor equal - weight, OLS, and XGB models had losses of 0.5%, a profit of 0.1%, and a loss of 1.3% respectively. The latest signals from the OLS model are bearish on all indices, while the XGB model is bearish on the SSE 50, CSI 300, and CSI 500 and bullish on the CSI 1000 [7]. 3.2 Treasury Bond Futures 3.2.1 Strategy Focus This Week - In terms of basis and cross - period spread, the basis of Treasury bond futures fluctuated narrowly, and the cross - period spread rebounded slightly. The capital interest rate continued to decline. The report suggests continuing to pay attention to the positive arbitrage and cross - period positive arbitrage strategies [60]. - For the futures timing strategy, the net value of the multi - factor timing strategy fluctuated this week. The strategy signals are mostly bullish, with main bullish factors including basis and high - frequency factors [60]. - For the futures cross - variety arbitrage strategy, the latest signals of the TS - T and T - TL strategies are bearish [60]. - For the credit bond neutral strategy, the current credit bond duration rotation and hedging strategy holds the 3 - 5 - year index with a longer duration in the cash bond and conducts Treasury bond futures hedging [60]. 3.3 Commodity CTA 3.3.1 Commodity Factor Performance - The domestic commodity market was differentiated last week. The decline of the US dollar index and the increasing expectation of the Fed's interest rate cut promoted the rise of precious metals, crude oil, and metal futures. The black - series commodities rose due to the rebound of coking coal prices and expected policy support. The performance of commodity factors was mixed, with price - volume trend factors slightly declining, term structure factors slightly rising, and basis and warehouse receipt factors slightly falling [80]. 3.3.2 Tracking Strategy Performance - The CWFT strategy had an annualized return of 10.0%, a Sharpe ratio of 1.69, a Calmar ratio of 1.13, and a maximum drawdown of - 8.81%. The return last week was 0.00%, and the return since this year was 2.93% [81]. - The C_frontnext & Short Trend strategy had an annualized return of 12.4%, a Sharpe ratio of 1.88, a Calmar ratio of 1.84, and a maximum drawdown of - 6.72%. The return last week was 0.69%, and the return since this year was 2.55% [81]. - Other strategies also had their respective performance indicators as detailed in the report [81].