Unrealized loss of -30 on freshly restarted dumb beta algo already, some mean reversion will likely be in order...
MFE since algo start Wednesday ~700, now back down to 300. NQ went unreasonably strong yesterday but the whole point is for me to not make discretionary trades unless super certain anymore, so abstained locking in $$$. Will see if I can make algos tweet their trades or something, kind of boring with a journal where I just say whenever I occasionally start or shut down an algo... (and not useful for self review).
Decent data Tue-Fri for NQ, will mostly be monitoring so that I don't around beta heavy in case of an unexpected down move. Besides that we have warm up for G20 that could cause extra random volatility that I can't duck. On dev front: 2.7 sharpe market neutral algo was probably due to a bug (not mine, in the platform!) and as it was fixed, my backtests started showing that it's worse than B&H. Also, the same bug impacted low res data in general making one of the algos I'm running live suffer worse DDs with lower Sharpe (0.75 instead of 1.1). Going to redesign market timing elements of that therefore.
Overslept data today as a small powernap turned longer. One algo down 0.2% other down 6% (yeah... eating up a large part of favorable excursion of equity in last two weeks for me, then again I am running this with rather heavy leverage on account I'm allowing to hit ESMA-protected zero on Black Swan, so it's well within backtest parameters).
Shit day to be leveraged dumb beta though even if leverage reduction did kick in to conserve capital as intended. -8% and -4% on the CFD account in total, wiping out gains since I started these in mid June. Just added a bunch to my unlevered long equity accounts, can see no particular reason for this price drop other than elevated buying frenzy last week and today's data being a reality check. Of course, the PA inspired will claim this is a triple top or whatever.
Fixed market neutral algo it would appear by adding a shorter time frame filter to it. Sharpe 3 (with heavy slippage modeled), here I come! (maybe, unless another bug pops up). Running on practice accounts for a few weeks just to confirm forward is better than backtest.
Down $600 on CFD accounts this week (MAE about $1100 or so). Back into slight profit for June again. We'll see how disappointed market gets when Trump and Xi achieve nothing on G20. -.- Market neutral algo hasn't made any trades yet (it averages 20 pair trades per month) but can go a week or more of zero trades in backtest, so I'm impatiently waiting for that. Want it to make about 15-20 trades so that I can see that forward matches backtest before launching on real $$$.