The HST model uses leveraged ETFs based on Nasdaq Composite ($TQQQ*) and on S&P500 ($UPRO) in its version 2.

[* this is the most popular ETF, but not the only one. Other issuers provide similar instruments, and some are traded in other markets, especially in the EU's countries; their behavior should be the same]

The calculations made by HST to determinate every trading day's gain/loss are made on the assumption that

• a 3x ETF is performing exactly 3x times the daily performance of the underlying index.
  Meaning that if Nasdaq 100 makes a +1.00% today, the ETF called $TQQQ will make a +3.00%.

This is oviuosly not true, there are approximations, every day at the closing, and during the day.
But these ETFs are so effective in many ways, that Hari Seldon Trader uses them largely: you can basically trade an entire index, an entire commodity, an entire sector, with low charges and quite precise performance.

How precise? Since we use $TQQQ a lot (HST model has been Aggressive Long = virtually holding $TQQQ for the 62% of the trading days in last 49 years), it would be nice to see if the real performance of $TQQQ is actually copying the move of the underlying index.

Nasdaq index was born in 1971, while $TQQQ inception date is February 2010. So we have 10 years to compare, and here is the outcome:

Surprisingly the actual performance of the real ETF $TQQQ has been higher than the virtual 3x index.

• $TQQQ               average annual yield 2010-2020 = 45.54%
• Nasdaq (daily 3x) average annual yield 2010-2020 = 37.95%
  
  We think that this is because the decade 2010-2020 has been particularely positive, going througha very long bullish market, with few, small drawdowns.
  There was no significant bear markets as in 2001 or 2008-2009.
  We fear that during bear markets the ETF could magnify losses instead of gains.
  
  

Still we have found a very good correlation between ideal and real, and this is a confirmation for our investment targets.

---------------------------------------------------

Update, July 18th 2020

HST has been Aggressive Long since the evening of April the 8th. At that time our members were suggested to abandon the CASH position, and to

• a: BUY $TQQQ in case they apply variant 2, or
• b: BUY $QQQ in case they apply variant 5

In our calculation (made on excel files), we consider that the ETFs are exactly copying the performance of Nasdaq Composite,
with 1x in case of Variant 5
with 3x in case of Variant 2

but what has been the REAL performances of $QQQ and $TQQQ?

Let's see:
Since the evening of April 8th the THEORETICAL PERFORMANCES would have been:

+29.81% for the Variant 5
+107.05% for the Variant 2

The REAL PERFORMANCES of the actual ETFs have been:

+29.34% for $QQQ (from $200.57 to $259.42)
+105.32% for $TQQQ (from $54.47 to $111.84)

CONCLUSION:
Differences from our VIRTUAL ETFs (calculated multiplying the performance of the index) and the REAL ETFs ($QQQ and $TQQQ) are really low.