I think I've got some 'leads' to an edge for trading (I only look at commodities and ES). I'm wondering if there is some place I can pay to learn how to backtest, as I have no coding/SQL or any of that experience. Right now the edge I believe I have definitely is not an intraday kind of thing but more so an 'it's bound to go in-the-money eventually, and I just need to keep constraining to get down a maximized risk/reward'. For instance, on silver chart, it printed not too long ago on the weekly a very bullish hammer candle with a long tail wick. Looking at chart history and doing rudimentary backtesting, I can see that every time such a candle printed (on the weekly but also this should translate to other timeframe charts) with a long wick, eventually it had a 'sizable' move in the direction one would think (down if it was a bearish wick with long high side). Not every time, of course, but in general very high likelihood. I also noticed on silver just using Investing.com that every time it got on a run of 5 or 6 straight green candles on the daily, it went relatively soon to test the 200 and other important MA lines. That just happened with silver obviously yesterday. I also had tons of conviction because I just look at the inflation that's been created and think this is market makers and the London fixers trying to get the weak longs out of PMs with this major run down to keep the Potemkin stock mkt village propped up. I also like integrating long term (25 yr or so even) charts and drawing the trend lines that most broadly encapsulate, and if it is say a ton of % away from the high side of a downtrend line, and close to the support, looking to just get long and be willing to sit in a position (this is VERY long term trading--it can be anyway- as coffee recently offered this opportunity and would have worked out great already) but I still see it as just a keep it simple stupid kind of way to look at trading. Yes it could take very long, but I'm willing to wait. Same with say buying SH or some inverse ETF on ES and just waiting to be ITM sizably eventually as rarely does ES blitz through so high without having retracements back through an area, and you can wait for when it does get overextended via numerous measures like RSI, etc. Just wondering how hard it is to learn to code and I'd like to get a signal that alerts me of such optimal setups, and I'm sure this could translate to say 5 min or hourly charts. These strategies are pretty unrefined as I described, and could require taking some heat, but I would use constraints and something like SOLVER on Excel to maximize the right risk/reward and have a good exit point versus the other thing I was mentioning where i just stay long and eventually I'll be ITM (dare it to do something extreme).
You should scroll down further here at the forum to the Programming section @ http://www.elitetrader.com/vb/forumdisplay.php?f=65 I'm sure if you just post a question about how to program...someone will point you to books, threads or something that can help you to learn how to program.
That is a simple code to program in easylanguage. if open > close then if close > low * 1.1 then buy next bar at open; *for a tail 10% lower than the bottom of the candle, for example. My main experience is with tradestation 9.0 and I was able to do hundreds of iterations (manually) in a matter of weeks, and eventually found my Grail Algo, as you are also seeking.
I think what I've come up with at this point is doable, but it also is not conventional. It is not about higher highs or anything remotely along those lines. I just have noticed some things about what happens when a long wick is printed, and so I'm thinking I need--for backtesting--years' worth of 15, 30, 60 min candle data (need OHLC of each candle of this whole duration) and to somehow pinpoint it to as you indicated basically, if " ..." buy or sell (depending on if the bull or bear kind of wick is printed). And from there the position hypothetically opened only closes when SL or profit target is reached, and then it's just waiting for next one. And I need to see what happens in aggregate and ideally in different times frames as well. I think it is conducive to crude in how crude--while volatile-- trades within ranges at least and rarely has something like a waterfall from 145 to 33 awhile ago that it did.
Tradestation can do all that. You get years and years of data access included with the monthly fee. I backtested 13 years of hourly bars on SPY when tuning my bot.
I'll probably call TS and maybe IB and see if they would have that as a package as you are saying. From getting the data exported, though, I still have the issue of not being very slick with Excel or anything coding wise. Getting all the OHLC on 15 30 60 etc. min charts for years would be great to have in Excel, but then how to somehow select just the implicit candles that were of the type I'm looking for and from there run the test of what would happen SL or tgt profit hit, not sure there. Hence why I was talking about looking for someone who knows how to do this and would be willing to teach lessons or something, unless something like you were saying of dictating what I want and having a system or someone translate it into code speak can be done.
just did manual backtesting again for copper, 15 min chart, 2 weeks' worth of data. 20 total trade setups I had (1 to 3 of them overlap so I'd have to literally go from start to finish, but still think it's fine for now). 9 ticks is profit tgt, 6 ticks is SL each trade. 12 clear winning trades, 6 clear losing ones, other 2 were on the fence, but I'd say 7 total losses. Actually my 'ticks' are in .001 when the tick size on the futures is .0005 for copper, so everything multiplied by two and calculated, with 12.5 tick value, it would be $1650 profit w/o factoring in commissions for the 19-20 trades. This is over two weeks, assuming one lot entry each time. All of this completely off hand literally writing down what happened after the candle I am looking for printed. Good winner/loser ratio of 9:7 and winners 50% more ticks than losers, so seems like a start for sure. But two weeks' worth is not near enough. And I just off hand chose 9 and 6 as win and SL after looking over the results for each of the 20 setups, so that is not optimized at all and some of these were to 15-20 tick winners if could take some more heat, although I tend to think this general ratio I got for what I did would hold as winners to losers would not be as favorable. I also like more minimal drawdowns at this point for acct in terms of risk mgmt/capital preservation.
Most cities and all over online have free Excel classes/seminars. Just use Google to find something and even Office Microsoft gives free classes in Excel. In fact, its so easily available that I'm now wondering why you have not use such a rich resource for learning how to use Excel unless your thread is really about you trying to find someone to do specific candlestick pattern coding for you instead of you learning how to code it yourself.
I'm not as concerned with the Excel part as I am just finding the data first off (OHLC for 15, 30, 60 min candles) going back at least 10 years so I can legitimately backtest this, and then from there highlighting the candles that I'm looking for that come up not even 1% of the time, and seeing what would happen upon bull/bear entry whether it'd hit my SL or profit tgt each time, and see what happens in aggregate. I presume there is nowhere on the internet for free one can get that long back of data, and that it will have to be paid for, correct? My best bet is probably to talk to IB or TS and see if they are able to somehow do the backtesting or know people who do. All I've done at this point is manual backtesting and I'd like to refine and optimize and do more testing.
TS from their website definitely has the data. Novice question, but generally if I just want to trade one lot maybe two at most at any one time in copper and crude on NYMEX, will I need to maintain about 4k in my acct. at all times? Just wondering what is tolerance until a margin call? And whether this changes if hold position overnight, even if stops are in place so no big risk. I'm looking at 15 min chart setups, so they take some time to play out usually.