Hello Graig, > Another question: I downloaded all the Nasdaq stocks going back ten years, but it seems > I'm going to have a severe survivor bias because the only stocks it downloaded were the > ones that survived the whole tens years (true? - it seemed that half the stocks got 404 > errors in AmiQuote). I got the Nasdaq ticker list from the AmiBroker website. >
If you mean the Amibroker.com site the list might be old - there is a post with some more recent lists at the UKB - refer to Data>>Yahoo>>DatabaseManagement "Setup A Custom Database - Nasdaq" (files attached to the end of the post) http://www.amibroker.org/userkb/ (the post is due for some maintenance as per HaiBoon Chan's comments - haven't got around2it yet). I believe that anyone who uses Jim's US-Stocks database, and keeps updating it, will eventually build up a database with redundant stocks included (you have to wait a lot of years though). My cheap trick, to get around survivor bias, is to pick out all the current declining stocks and see how my bull trades handle that - also the performance of the reverse (bear trade) on the upwardly biased stocks is revealing. > Another strange thing I just noticed is that 99% of the trades are with extremely low priced > stocks (e.g. .02, .04...). Hmm... That's the nature of the Nasdaq (it still seems very popular with sections of the trading community). Early on I found that low priced stocks unduly biased the results of certain types of systems e.g. a 0.02 stock will move 50% on a 0.01 cent move (and they don't come any smaller than that). Also, because they are lightly traded the stock might sit at 0.02 for days and then when a customer turns up move 0.03 in one step, on light volume, (it can make,say, a breakout system look very good - in theory). IMO - (for equity trading) - to maximise returns you have to leverage eventually (once wins in the underlying are consistent) - for that reason I gravitated towards the high vol stock where the leveraged instruments are anyway (stocks that are covered by options in the US and CFD's outside of the US). I believe signals are more reliable in high liquidity stocks. Just a thought - say a trader only ever followed stocks that are in the Dow and stopped trading them as they fell out of the index - do you think the Dow included stocks would be subject to large survivor bias (over the long term)? The same question applies to the component stocks of any major and small index. brian_z --- In [email protected], "cmaiman" <[EMAIL PROTECTED]> wrote: > > Hi All, > > New to AmiBroker and had a couple of questions. > > To get my feet wet I'm trying some code from Howard's book and am not sure I > understand the results. > > Below is an example of two trades using the code in Fig. 22.22 on Pg. 315 (I did optimize > it as an experiment). My question is, are these realistic profits? It seems to me that it > doesn't take into account bid/ask and that the likely actual sale price is lower, perhaps low > enough to be a loss. If it doesn't take bid/ask into account, how do you do that in > AmiBroker? > > Ticker Trade Date Price Ex. date Ex. Price % chg Profit % Profit > -------------------------------------------------------------------- > AAUK Long 3/23/1998 3.43 3/25/1998 3.6 4.96% 430.20 4.90% > PTRO Long 11/17/1998 0.28 11/30/1998 0.38 35.71% 3566.43 35.66% > > Another question: I downloaded all the Nasdaq stocks going back ten years, but it seems > I'm going to have a severe survivor bias because the only stocks it downloaded were the > ones that survived the whole tens years (true? - it seemed that half the stocks got 404 > errors in AmiQuote). I got the Nasdaq ticker list from the AmiBroker website. > > Another strange thing I just noticed is that 99% of the trades are with extremely low priced > stocks (e.g. .02, .04...). Hmm... > > Thanks! > > Craig >
