Daily performance today was flat.. started last night at 6pm at 28,476 and closed at 28,406. Real loss would have been higher due to high activity level. Broker would be happy. 90RTs in EURUSD alone. USDJPY was the losing side today. If people want can post full trade list too.
interesting you ask .. i used to concern myself with that info and i know a lot of folks wont agree but I have found stats pretty useless when algo trading. my number one factor i consider is what is my slippage real versus machine and then how can I improve. I use a process that is consistent over time, with average trade levels in the $35-45 range. dont look at ratios because to me they mean very little. today was a good example. we jumped 1100 at 7-8 pm then dropped 1400 but then jumped 500 and settle back to flat. it was due to behaviours beyond my design and how i think the market behave "generally" over time
sorry there is another number you quoted that i do consider and its important to me and IMO all traders.. its % right.. whenever i see this getting above 50% i start to get worried and above 60% right I wont trust the system. Most traders that seek high% profitable don't understand that that is the domain of high end Stat ARB HFT but in general a falicy. trying to be right all the time in variably leads to poor risk management and eventual kaboom
not sure about that.. i would have agreed with you in the past but I can a 35 man trading group using gray box and was dumbfounded by how poor returns were.. does this mean you will join?
yes i do but they need to come out of excel. I use the tradestation strategy as starting point. cant build a real program in TS. then run algos on equity curve and then I have an auto trade bridge. The models generate approx $1400/$1800 per month real with risk approx $2500-3000 but recovery is few days.. these are more important numbers than sortinos imo. fyi the numbers will be much higher in this demo world but again this is learning exercise and want to get other ideas