2022年美赛（MCM）C题M奖论文

inconsistent format of Date column programmatically and sorted it, then we determined the

The approximate trend of gold and bitcoin prices from September 11, 2016 onwards is

depicted by plotting a line graph.

For the first question, considering that the csv file only gives the relationship between

did a 1st order difference and the data passed the stability test. To avoid errors caused by

model. Then we give a perturbation to the trading strategy to see if the total return will

For the third question, we were asked to perform a sensitivity test on the rate parameters.

to 1-5 times and 1-3 times their original values, and obtained

, α and β of the ARIMA model in the first question separately, and compared the predicted

transaction Timing Theory VaR model

1. Introduction....................................................................................................................2

4. Establishment and Result of the Model.........................................................................6

4.1 Task One..................................................................................................................6

4.1.2 Analysis of the Problem...................................................................................8

4.1.3 Model Establishment........................................................................................8

4.1.4 Algorithm Design ...........................................................................................11

4.1.5 Solving the Problem........................................................................................12

4.2 Task Two................................................................................................................13

4.2.2 Analysis of the Problem.................................................................................13

5.3 Sensitivity analysis of the three parameters of the ARMIA model........................15

6. Model Expansion...........................................................................................................16

6.1 Quantitative transaction Timing Theory Approach................................................16

6.1.2 KDJ Indicators................................................................................................17

6.2 VaR Risk Control Model........................................................................................18

7. Strength and Weakness .................................................................................................19

7.1 The Strength of the model.......................................................................................19

7.2. The Weakness of the model...................................................................................19

8. Conclusion.....................................................................................................................20

Quantitative transaction was created after the 1950s and has been developed for nearly

70 years. It has developed more maturely around the world and has gradually become popular

in the capitalist markets of major countries[1].

Quantitative transaction refers to the use of computers to build mathematical models to

judge the timing of trading for buy and sell operations. While traditional securities trading

model with the gold and bitcoin trading data in our hand and calculate the profit outcome of

an initial $1,000 investment on September 10, 2021.

The second question asks us to argue that the model provides the best strategy.

The third question asks us to perform a sensitivity analysis to indicate the impact that

transaction costs have on the quantitative trading strategy and results we have developed.

The final question is asking us to write a two-page memo including the strategy, model,

and results to our investors.

This problem requires us to develop a model that provides the best trading strategy.

Such a model is known as a quantitative trading model in the field of finance and is one of

limited the data parameters to only two labels, time and price. Obviously, it is not feasible to

apply common models and algorithms here.

plotting the line graphs of the trend of gold and bitcoin prices from September 11, 2016, and

comparing the line graphs in the title overview, we found roughly the same, which shows that

our data cleaning efforts were successful.

In the second step, we proceeded to solve the first half of the first question, which is to

predict the price after a certain day from the original price time series and thus calculate the

return of the latter days. After thinking about it, we found that only time and price were given

to get the new time series, and then performed the smoothness test we found that the new

predict the price for the days after the closing, which is used to calculate the return for each

positions in the short term.

In the third step, we argue for the optimality of the model as required. In fact, the

ARIMA model has inherent shortcomings, which we list in Section 7. Second, the data in the

In the fourth step, we performed sensitivity analysis on the gold and bitcoin trading rate

to 1-5 times and 1-3 times of the original, and obtained little

, α and β of the ARIMA model in the first question separately, and compared the predicted