We take the good with the bad.  An extremely difficult two weeks to say the least, though as stated before expect moves from 4,000 to 12,000 in positive and negative moves.  The algorithm performed as expected, but operational errors cost us. Drawdowns of this type are normal, as experienced during last year (3 times through the course of the year) but drawdowns outside the scope of the algorithm are not expected.  Training is being provided so that there is plenty of back-up in case one cannot physically be at the terminal to turn systems on and off.

The good news is tracking, what the model did, had us at +$300, but actual transactions had a very ugly drawdown.  This operational risk has been resolved, and something that is non-technical and will not re-occur.  Overall, algorithms performed as expected. Additionally, we did not have 3x risk on all year, and as such we would be up nearly 7% for the year even with this setback.

Emotional trading and a snapback rally from year lows caught us long as the market sold off (a hedge on where it was not requried) and a naked short during the rally (when a hedge was requried).  1-2% moves overnight on very light volume kept the market trending higher as we eventually broke 1300.

THERE WILL BE A CONFERENCE CALL APRIL 3RD, TO DISCUSS THE SYSTEM, ITS RISKS AND BENEFITS.   

 So, what happened over these last couple weeks?  Just to be clear, you should expect a 25% drawdown to occur at somepoint in the trading system, this is normal.  Additionally, we have ramped up signifcantly the leverage (3x the leverage from last year and this is where we need to be as we were under leveraged last year).

What happened was not so much an issue with the algorithm, but with operations.  A long position was left on when it should have been removed, and a week later the long position needed to be put back on to protect upside risk.  I have started training some individuals to assist the algorithm when I cannot physically be there to "turn" systems on and off.  I covered the losses, since this was a human error, but the percentage drawdown will remain since this is a "risk" though it never hurt as much last year when it happened.

Just to clarify, this did happen several times last year as part of the natural occurance in the algorithm's life.  We were short when the ECB announced a 1 trillion dollar bailout, this cost us significantly (6% with last year's leverasge and an 18% drawdown with this year's leverage).  In fact, the first day of trading cost us a $1200 loss (a 3% drawdown last year's leverage, and a 9% drawdown this year). 

The last risk item that I cannot control at all is working capital, leverage risk.  If we move our leverage by 3 times, and we sustain a loss the very day we apply the new leverage, there is little you can do since we have not had 3x leverage for the full year.  If we had 3x for the full year we are actually UP nearly 7%, even with this 8% drawdown!

So, in reality, the algorithms held up just fine and an 8% drawdown is normal (in last year's terms this is only a 2.6% drawdown).  Additionally, in order to handle the kind of account we want, we needed to get through trading in an extremely volatile environment with multiple contracts, and I think we survived just fine and demonstrated resilience under adverse conditions (everything that could have gone wrong, did!).  Overall, we survived and that is what I had really been wanting to test.  I can't say we're out of the woods on this drawdown, but at least it won't be because of human error again.

So, can we still hit 30% this year?  Its still quite possible, I am testing the options system, and one of the plays in the options system is when a hedge goes bad.  In other words, we have covered the $9,000 loss with a set of covered calls, so in effect we didn't experience a loss (yet), but I am not counting this in the system as of right now.  This strategy is to "rent" a position out by taking in a 1375 Call Premium, the issue is that with this strategy it does potentially lock in the 9,000 loss to be incurred at a later date.