EminTrader, The Buy/Sell prices for 4/21 and 4/22 trades for HCBK do not jive with stipulated L5 %, or the Open price values for those days. There are definitely some computational errors for those 2 days.
please be advised the current market situation is quite different from normal. i actually doubt a fixed L5 would work for long term since if that is so simple, every one should have made money. I do not think that this approach has not visited by so many quants there on the street.
I compared my backtest result with emintrade's realtime trading record, and tried to find out why my backtest did not perform as well as his real time trading. I found some interesting results. the most difference comes from the opening price. it seeems the openning price from yahoo and esignal is quite different from emintrade's openning price. second, for some cases, my backtests were stopped out while emintrade recorded a decent gain. for example, on 2009-02-25, for stock RIMM, the L5 is 1.7581%. the result from emintrade is: RIMM : Buy at 41.365. Close at 42.079 So + 1.7278 % based on his entry price, i calculated the opening price=41.365/(1+0.017581)=40.65 and the entry price would be filled during 9:45 to 10:00 am. after 10:00 am, the stock traded down to the lowest 40.16 around 12:45pm to hit the stop price, and then traded up to hit the profit-taking price at 42.08. I wonder why emintrade did not record this trade as a loser? in my backtest, the open price is 40.72 from both esignal and yahoo. so many cases his opening price is different from yahoo and esignal. I began to question if emintrade really made those trades.
You are right...I made a mistake on 2009-02-25 for RIMM. I use prorealtime. Opening price : 40,65 -> Daily datas -> i use it for opening price. It is public. They are free. 40,72 -> Intraday datas...They made also a mistake. Perhaps i was not clear : I trade my emintrader trading system in real but since i am posting in forum my L5% values i do not trade these 5 stocks. I use others. Results : HCBK : Buy at 12.844. Close at 12.81 So - 0.233% MRVL : Buy at 10.444. Close at 10.41 So - 0.325% RIMM : None JNPR : Buy at 21.787. Close at 22.273 So + 2.2334% SCHW : None 2009-04-27 : HCBK : L5 = 2,0992% MRVL: L5 = 2,9668% RIMM : L5 = 2,493% JNPR : L5 = 2,2844% SCHW : L5 = 2,2117% Have good trades... Overall %result : + 133,28% FSLR: L5 = 2,8512% SKF : L5 = 4,4556%
Another anomoly - It seems like L5 changed for HCBK on 4/1/2009 from 2.0922 to 2.0561, but hasn't been reflected accurately since this date.
No....it was NEVER clear that you were not trading the stocks you provided the L5% values....the irony is that they have performed brilliantly. So, I have to ask the question are the stocks you supposedly trade better or worse that the 5 you listed? Why not show the stocks that you trade? It really will be of little significance ( price performance) in the market place and will certainly enhance your credibility to this really good thread. I am a little disappointed, but did suspect that something was amiss a few pages back when you had to make a correction to a "trade" that would have otherwise been obvcious to you on your trading account. Please don't take this as criticism...you simply caught me off guard. NiN
Since I started spreadsheeting this stuff, I've found too many mistakes. I wondered how these mistakes could be made, if you were trading from them. Now I know. Why would you post these, if you are not trading them ? What is the motive here ? Hmmmmm......
this would be my guess: 1) publicly document a system he developed 2) share some of his insights with the ET community I tend to always think the best of people even though been burned quite a few times . You have to admit this is pretty rare , someone calling calling entrys and exits ahead of time . All easily verifiable .
Thank you babe714, Perhaps that hayman want to see a trading system with an automatic backesting. I attached in this post this kind of system : 18 years of backesting !!!!, included Commissions !!!!, ridiculous drawdowns !!! results are given by computer. Initial capital of 10 000$ -> more than 450 000$ in final Ok i do some mistakes (some because of prorealtime and others by my fault ; some are without impact on results and others yes) but it is as trading in real...I will try to be more concentred on my elitetrader journal. I change formula : i will give you my L5% values each day but i will add just 2 informations : opening price (clear ! -> proof on prorealtime, it is free) and result per stock. I do not trade these 5 stocks because i don't know if my system will be as good as if 1000 people follow it ; even if i believe it will neveer arrive in the future. I have choice. Some stocks performs better than others. But it is based on past results.