3X ETF Short Hedge Algorithm

Impressive! Are there any indications on IB for the cost of borroow on these? Also, have you tried different rebalancing periods, out of interest? I'm thinking a lower rebalance period would potentially yield more stable results, but I'm not sure.

IB's borrow cost starts at 0.25%, their statistics say the average is 0.3% but can go up to 30% during a short squeeze. They don't have statistics for individual stocks but I don't worry too much about it, and I'm short more often than long.

The 3X ETF short strategy is somewhat sensitive to starting and rebalancing periods. If you happen to rebalance when the ETF pair is out of sync (and you are at a loss), when they eventually sync again you will lose even more. But if you hold the pair throughout the correction you'll get your money back and more.

I would be wary of tuning the rebalancing frequency as it's literally the only parameter of the strategy and you'll likely introduce a lot of overfitting bias into the backtest.

What instrument you are using for shorting ? Is it options ?

The algorithm uses the Inverse version of the ETFs to short. Look up the tickers "NUGT","DGAZ","UGAZ" to read more.

Shorting these ETFs tend to trigger frequent Hard to Borrow interest penalties which can make shorting very costly. Correct me if I am wrong but it seems these particular ETFs were cherry picked… other 3x ETF pairs don’t yield very good results… http://www.3xetf.com/  try the 3x Crude ETFs OILU/OILD

Hi Yujun,

Thanks for presenting this to the community. I am trying this now with slightly out of the money PUT options with over 100 days to expire on my IRA account. I'll probably be in the red for a few days because the market is moving sideways.  But im hoping that this proves fruitful since the backtests performed so well. 

Longer period is not so great.

"3X leveraged ETFs always decay over time. For example, if you have an ETF that starts at 100%, goes up 25% to 125% and then goes back down 20% to 100% you are back at square one. But the 3X ETF will go up 75% and go down 60%, ending up at 70%."

I don't know if I'm allowed to revive this, but this is just faulty b.s. logic. 3x commodity ETFs trend down because of contango/backwardation but volatility drag, if you actually crunch the numbers is not a big enough factor to cause 3x stock etfs to trend down as long as the underlying assets trend up. The reason your strategy is working is because you're cherrypicking ETFs that inversely correlate with the market or commodity ETFs.

I have a similar stratgy that uses long/short pairs and rebalances every day - I used it live in Quantpian but havent tried to rebuild it here - had very good results

def initialize(context):
    """
    Called once at the start of the algorithm.
    """   
    
    # Rebalance every day, 1 hour after market open.
    schedule_function(rebalance, date_rules.every_day(), time_rules.market_open(hours=1.5))
     
    # Record tracking variables at the end of each day.
    schedule_function(record_vars, date_rules.every_day(), time_rules.market_close())
    
    # IB parameters
    set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.00))
    
    context.leverage = -1.0 # short both pairs
    
    context.etfs_long=[ sid(37049), sid(37515),  sid(41969), sid(42470), sid(39214),sid(45570)]
                      #  FAS, TNA, UVXY, DGAZ, TQQQ, JDST
                       
    context.etfs_short=[  sid(37048), sid(37133), sid(41968),sid(39211), sid(42477), sid(45571)]
                      #  FAZ, TZA, SVXY, UGAZ, SQQQ, JNUG   
        
    context.sim = -1 # change to index of any pair obove (0 - 4) to trade that single pair
         
     
def compute_weights(context, data):
    """
    Assign weights to securities that we want to order.
    """
    num = 0
    
    for (l,s) in zip(context.etfs_long, context.etfs_short):
        if data.can_trade(l) and data.can_trade(s):
            num+=1
            
    if (num>0):
        return context.leverage/(num*2) # we have pairs so half for each
    else:
        return 0
 
 
def rebalance(context,data):
    """
    Execute orders according to our schedule_function() timing. 
    """
    if context.sim >= 0 : # trade single pair
        if (data.can_trade(context.etfs_long[context.sim]) and data.can_trade(context.etfs_short[context.sim])):
            order_target_percent(context.etfs_long[context.sim], context.leverage/2)
            order_target_percent(context.etfs_short[context.sim],context.leverage/2)
            
    else: # trade all as a basket
        weight=compute_weights(context, data)
        
        if weight:
            for (l,s) in zip(context.etfs_long, context.etfs_short):
                if data.can_trade(l) and data.can_trade(s):
                    order_target_percent(l, weight) 
                    order_target_percent(s, weight)
        
 
def record_vars(context, data):
    """
    Plot variables at the end of each day.
    """
    
    record(leverage=context.account.leverage)
    record(positions=len(context.portfolio.positions))
 
 
def handle_data(context,data):
    """
    Called every minute.
    """
    pass

Same idea except a daily rebalnace and 6 3x long and opposite 3x short ETFs 

Sharpe Ratio

2.858

this has some larger drawdowns, think if a dual momentum type filter is added to stop trading if cash > a bond fund go into cash.