Yes, you could use the record_data() function. You'd likely need a bit of math and store variables in context.
Pseudocode:

initialize():  
  context.portfolio1_return = 1.00  
  context.portfolio2_return = 1.00

  context.tsla_last_px = get_last_price('TSLA')  
  context.spy_last_px = get_last_price('SPY')

  schedule_function(record_data, EVERY_DAY_AT_CLOSE)

record_data():  
  tsla_return = get_last_price('TSLA') / context.tsla_last_px  
  spy_return = get_last_price('SPY') / context.spy_last_px

  context.portfolio1_return *= (tsla_return)  
  context.portfolio2_return *= (0.5*tsla_return + 0.5*spy_return)

  log_data(context.portfolio1_return)  
  log_data(context.portfolio2_return)  

Another approach is to fetch the backtest results in a notebook then compare inside that notebook.

One can retrieve all the data stored in a backtest, including the daily performance, daily portfolio value, and recorded variables, through attributes of the AlgorithmResult object. This is all done in the notebook environment. The first step is to get a reference to each backtest using the get_backtest method and the backtest ID. The backtest ID can be found by opening a backtest and then copying the last string of digit after the last forward slash. Something like this.

# First get the two backtest results  
backtest_id_1 = '5e528e3c9e2bfd435c19cefc'  
backtest_1 = get_backtest(backtest_id_1)

backtest_id_2 = '5e528dc21ff9b7434f987de4'  
backtest_2 = get_backtest(backtest_id_2)

The returned AlgorithmResult object has a number of attributes which fetch different data about the backtest. There is a complete list of these attributes in the documentation (see https://www.quantopian.com/docs/api-reference/research-api-reference#algorithmresult-attributes). The cumulative_performance.ending_portfolio_value attribute will fetch the daily portfolio values. These could then be used to compare our two backtests.

backtest_1_daily_values = backtest_1.cumulative_performance.ending_portfolio_value  
backtest_2_daily_values = backtest_2.cumulative_performance.ending_portfolio_value

To plot these results on the same graph perhaps use the pandas plot method. First get them into a single dataframe, then plot. There are of course a lot of ways to plot but this is convenient.

# First make an empty dataframe  
plot_df = pd.DataFrame()

# Next add the daily values as columns  
plot_df['backtest_1_value'] = backtest_1_daily_values  
plot_df['backtest_2_value'] = backtest_2_daily_values

# Finally, simply plot the dataframe  
plot_df.plot();

Attached is a notebook which includes the above code. Make sure to replace the backtest IDs with your own before running. The notebook will NOT run with the current IDs since backtests are private (ie nobody other than the account holder has access).

Hope that helps.

@Harry Sauers and @Dan Whitnable

Thank you very much for your help. Both are great ways of solving this problem!