Although August shows it down 7% that is only and end-of-month figure. It is possible they may have been down 30% at one point. Nevermind. This actually seems 'kinda' safe. If you have a $40k account and you're aiming for 1% monthy returns, you need to sell 2 of the weekly ES contracts at 2025 (delta 5) $100 per week x 4 weeks ~1% for a $40k account. If it drops 5% in a week to 1985 you lose about $3800. A 10% loss is bad but not blow up.
It's unlikely he drew down 30percent. More likely he drew down 3percent twice. Considering the amount of vol during that time, his returns are good. At the end of the day it's short vol; he won't profit in vol spikes. But he's proven to be pretty adept at navigating them. He gets the pnl of being short vol in quiet times while significantly reducing the asymetric payout (at least empirically). he doesn't blindly sell wings like the typical elite trader. He manages his delta and his gamma. And has prediction models to tell him when to get out.
Call me stupid, but I haven't yet seen the light on the usefulness of greeks. Considering that they are merely first and second derivatives of the BS model, from where does the prediction element come from when option prices are time varying?
You're right, the greeks are descriptive and offer no inherent advantage just by understanding them. There are two uses imo - the first is knowing the portfolio's risk and the second is allowing traders to tailor exposure to their prediction. Risk is easy for stocks, if I'm long $100K and the market moves up 10% I've made $10K. If I'm long $100K notional of ATM calls and the market moves up 10% it's unclear what I've made...it will depend on what greeks I had exposure to (not a great example, but let's go with it anyway). ...the prediction element is tied into the risk part. If I think realized vol will be less than implied I want exposure to lots of gamma. If I think the market will crash like 1987 I want lots of exposure to delta. If I think implied vols cannot go any higher and will collapse I want lots of exposure to vega. If the risk from the greeks doesn't match the prediction then it's unlikely the forecast will make money. As you can imagine, it's usually some combination of the greeks. The idea is an options position when properly tailored to match a prediction can outperform the risk adjusted return from simply investing in the underlying...that knife cuts both ways obviously.
I was curious too so I did a web search and found this information according to SJ Options: Hope Advisors' (her fund?) AUM was ~$225M around April 2013 (reporting period) and was around ~$250M in April 2015 (reporting period). Don't know about returns or withdrawals though. As a comparison, S&P 500 returned 16.7% excluding dividends during the same period and about 19% with dividends reinvestment.