Dr K UVXY Model

Dr K UVXY: After rigorous ongoing research into UVXY in the Dr. K laboratory, we will issue buy, cash, and sell signals on UVXY when the odds are deemed favorable. Since UVXY sometimes does not correlate with the CBOE Volatility Index (VIX), other variables were used in conjunction with the VIX such as institutional money flow in leading, high relative strength, super cap (>$20 billion) stocks.

There was a refinement made on 05-09-2012 which, while in beta testing mode, did not prove out, then further testing was done thus no signals were issued between 06-14-12 and 09-05-12, which culminated in the present version of the model as of the 09-05-2012 signal.

Note: Dr K UVXY is still a work in progress. The table below shows how the model has performed in real-time with the first three signals from 4-19-12 to 4-27-12 representing TVIX, and all subsequent signals representing UVXY, simply because UVXY has been found to inversely correlate more smoothly with market direction. The first signal is a cash signal, meaning to do nothing until the model switches to either a buy or a sell. As with all our services, members receive real-time updates via email whenever there is a change in signal.

Date	Signal	% gain / loss	$1 becomes(new)	$1 becomes(old)
04-24-2013	Sell	9.73	1.17*
04-24-2013	Cash		1.07
04-23-2013	Buy	-1.38	1.07
04-10-2013	Sell	6.29	1.09
04-03-2013	Cash		1.03
03-25-2013	Sell	-3.02	1.03
03-05-2013	Cash		1.06
02-28-2013	Buy	-8.54	1.06
02-27-2013	Cash		1.16
02-25-2013	Buy	-19.12	1.16
01-10-2013	Sell	12.83	1.44
12-26-2012	Cash		1.28
12-17-2012	Sell	-28.61	1.28
12-14-2012	Cash		1.79
12-12-2012	Sell	-11.74	1.79
11-14-2012	Cash		2.03
11-12-2012	Sell	10.72	2.03
11-07-2012	Buy	5.46	1.83
11-01-2012	Cash		1.74
10-18-2012	Buy	16.82	1.74
09-05-2012 ***	Sell	48.84	1.49
06-14-2012	Cash			0.53
06-14-2012	Buy	-9.32		0.53
06-11-2012	Cash			0.59
06-06-2012	Buy	-27.74		0.59
06-01-2012	Cash			0.81
05-31-2012	Buy	13.03		0.81
05-29-2012	Sell	-9.88		0.72
05-11-2012	Cash			0.8
05-09-2012 **	Buy	-3.76		0.8
05-07-2012	Cash			0.83
05-01-2012	Sell	-14.25		0.83
05-01-2012	Cash			0.97
04-27-2012	Buy	0.15		0.97
04-20-2012	Cash			0.97
04-19-2012	Buy	-2.53		0.97
There are no earlier results

*As of 05-17-2013 (intraday or end of day, depending)

** Reversion to the mean component and fail-safe level adjusted from this day forward

*** Further critical adjustments made to the model

Note, while the returns in the backtests resulted in MASSIVE gains in the high triple digit percentage range since inception of the TVIX (Nov 30, 2010) to the end of the backtest (April 10, 2012), a VERY HIGH RISK had to be expected. Indeed, losses as high as -38.2% on one signal were recorded, the largest drawdown over a series of signals was -56.6%, and larger drawdowns are possible in the future. To deal with this issue, a maximum loss fail-safe of -19% is in place which reduced the maximum drawdown over a series of signals to -33.6%. This means that on any given signal, the maximum loss should be roughly -19% while the maximum loss across a series of signals so far has been -33.6%. Of course, past testing never guarantees future performance as larger drawdowns are possible in the future

Nevertheless, it is ESSENTIAL that one position sizes with care when using these signals. We would suggest either using only a small amount of your capital, such as 20% maximum, or pyramiding a position starting with an even smaller amount of capital, then allowing yourself a larger maximum % only when you have sufficient gains in your position.

Since theoretical gains over a period of about a year in these backtests were in the high triple digit percentage range:

If you were to grow your 20% allocation by 50%, this would increase your total account value by a net of +10%.
If you were to grow your 20% allocation by 150%, this would increase your total account value by a net of +30%.
If you were to suffer a -56.6% drawdown on your 20% allocation, this would decrease your total account value by a net of -11.3%.

Note, to achieve the high triple digit percentage gains over the lifetime of the backtest, the -56.6% drawdown (not using the -19% failsafe) was made up in about 9 weeks. In other words, a gain of +76.7% (=1 divided by 0.566) was achieved over a series of signals in about 9 weeks, right after the -56.6% drawdown which occurred over about 6 weeks. Thus, signal changes can occur with greater frequency with the Dr K UVXY Model than with the Dr K Market Direction Model.

ALTERNATIVE to TVIX: UVXY is a 2-times vehicle that potentially inversely correlates more smoothly with market direction than 2-times TVIX so the table shown below will represent UVXY instead of TVIX after the first three signals (from 4-19-12 to 4-27-12). If one wishes to roughly halve losses, one may opt instead to use 1-times vehicles such as VXX (1-times VIX). VXX has been trading since January 30, 2009. In backtests, losses as high as -20.9% on one signal were recorded, and the largest drawdown over a series of signals was -31.9%, though larger drawdowns are possible in the future. Note that while losses are roughly halved, gains are also roughly halved. Thus your risk/reward stays the same but with roughly half the volatility.