EFTA00370224.pdf

From: Lesley Groff To: Jeffrey Epstein <[email protected]> Subject: Fwd: ATorus Daily Portfolio Report - 5/1 Date: Fri, 02 May 2014 15:51:38 +0000 Attachments: ATours_BacktestNAV_050114.pdf SSRN-id2308659.pdf Sent from my iPhone Begin forwarded message: From: Michael Fowler Date: May 2, 2014, 11:47:42 AM EDT To: Lesley Groff Subject: ATorus Daily Portfolio Report - 5/1 Lesley, Please find attached the Daily Portfolio Report for 5/1. Have a great weekend! - Daily Commentary - We wanted to share this paper with you: Pseudo-Mathematics & Financial Charlatanism: The Effects of Backtest Oveifitting on Out-of-Sample Performance (SSRN Paper and attached). This provides a good topic for how our methodology is different (at least we like to think so, and be intellectually honest), while also pointing out assumptions that we make. Our model is not 'adaptive' in the context of learning from out-of-sample to then applied to in-sample. Yet, at some level, out-of-sample is always the future. As we've stated, we don't use Feed Forward Neural Networks of Genetic Algorithms. Our equation in terms of dynamic reference frames is constant since inception (1898 in the case of DJI). To us "it just makes sense" that when volatility changes from your initial trade date you need to increase/decreases your sensitivity (there's a lot of subtlety in that statement when put into practice) to the current moment, given how we know how volatility scaling conditional on itself has constant exponents. This is not so say there are optimal and less optimal overall states of the system. This is why portfolio construction, our second pillar of the methodology (outside of the other two: (i) timing and (i) risk via position sizing and rebalancing ) is designed to ideally amortize out these assumptions by maximizing the randomness by "casting a broad net." Where there could be an issue of out/in-sample is in portfolio construction. We feel this has been mitigated given the constraints put on the opportunity set screener. Yet, this is where the risk lies to create an analogy pointed in the paper. Yet, given that the screener is dynamic, the opportunity set is conditional to the system evolution (market going up or down and including securities that no longer exist), eliminating selection bias. A quote from one of my favorite individuals stands out: EFTA00370224 "Another thing I must point out is that you cannot prove a vague thew), wrong. [...] Also, if the process of computing the consequences is indefinite, then with a little skill any experimental result can be made to look like the expected consequences." —Richard Feynman [1964] Best Regards. Michael J. Fowler The information contained in this electronic mail message is confidential information intended only for the use of the individual entity named above, and may be privileged. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution, or copying of this message is strictly prohibited. If you have received this communication in error, please immediately notify us by telephone, and delete the original message. EFTA00370225