sure they can be tested mechanically. the reason people don't is that when you do, it turns out they all suck. it's a discretionary tool, i think.
Kay, I'd be very interested in hearing you talk about how to do this more. From what I've learned so far about candlesticks is that testing them mechanically would be very difficult and have some judgement calls imbedded into any mechanical system. This is because you are: 1) recognizing patterns or groups of patterns that don't have precise mathematical definitions; 2) some of these patterns have to occur at certain times (e.g. the top of a trend) that also do not have precise mathematical definitions. This means two coders can backtest candlesticks and come up with valid but different results. Add to that: exit strategies, mm, intraday/daily/weekly/monthly periods, taking signals at the close vs O, H or L, market, limt, stop or stop limit orders, adding confirmation technical signals, and back testing gets even muddier from what I can tell. So how did you/do you come up with your conclusion that candlesticks can be mechanically tested and that they don't perform well?
it's very simple to backtest candlestick patterns. You look at the description of bearish engulfing. Then you describe the relationships between two candles' H, L, O and C. Like H1 less than O2 or whatever. Then if you wanna get picky, you can define the trend too. For an uptrend you can require some long-term stochastics to be over 80 or something like that. The exit is even simpler. Exit on close (or open) at +3, +5, +10, +15, etc. bars. If the % of winning trades isnt much higher than 50% then I'm sorry Mr. Nison, there's no edge.
First of all, How you explain the above in your quote to test candlesticks... I'm sorry...doing such WILL produce poor results...almost intentionally. The problem is your description or lack of a description for a trend along with throwing stochastics in there. No wonder your having problems with intraday candlesticks. (Note: Your reference may be in that of daily candlesticks...is so...please ignore this post) You can sweep it any way you want... we can be talking about candlesticks, macd, stochastics or fib numbers or whatever... if you exclusively use them by themselves...the odds aren't very good will be trading beyond a year. However...the key is to use candlesticks in combination with simple chart pattern recognition (price action alone patterns) or in combo with indicators. Note: Candlesticks aren't very reliable during choppy market environments...such environments the Stochastic does well. Traderkay...you should try using candlesticks with a trend like indicator such as the MACD or similar... your results with candlesticks will improve dramatically in comparison to using it with Stochastics...possibly the reason why your having trouble with candlesticks. Simply, candlesticks are very reliable in in during trends as continuation, exhaustion or reversal patterns. My observation is based on using intraday candlesticks on the Eminis for many many years...1min and 3min charts Once again...if your talking about daily candlesticks...ignore this post. P.S. You show me a trader that loses money via candlestick analysis on intraday charts... I'll prove to you that the trader was applying intraday candlesticks incorrectly. There are lots of traders here at ET successfully using intraday candlesticks besides me... try asking once what is it that your doing wrong instead of posting it doesn't work without any proof and you may get a private message with useful trading tips about how to effectively use intraday candlesticks. Review this post traderkay...if your interested...I will more than happily explain (go into details via graph examples) along with showing you in realtime during trading conditons (the two go hand in hand) how to effectively apply intraday candlesticks. http://www.elitetrader.com/vb/showthread.php?s=&postid=216063#post216063 Also...if my memory serves me correctly...I offered to help you a long time ago about intraday candlesticks...you didn't respond then unless I got you mixed up with someone else. NihabaAshi
I only really take notice if the daily chart puts in a reversal candle such as a Hammer or hangin man. Like the following link will show you that March 12th's strong reversal into the close forming a hammer on the daily chart was a tip off to this rally we have. http://www.elitetrader.com/vb/showthread.php?s=&postid=218367#post218367
Traderkay As usual...no reply from you after I responded to your statements... if your statements were in reference to intraday candlesticks. As I mentioned before...intraday candlesticks along with a decent trend indicator or price action alone pattern can be very profitable. Now...since you boldly used the Bearish Engulfing Japanese Candlestick pattern as an example (your choice...not mine)... lets take a closer look at the Bearish Engulfing candlestick pattern.. Since last March 10th Monday...I've posted every Bearish Engulfing JC pattern that has appeared on intraday charts (1min and/or 3min charts). Yep...its only a small sample...but if I had time...I would post every trading day for the past 2 years (I do have backtesting data on this particular pattern and several others) and it is an extremely reliable, repetitive and profitable pattern out of many. (Note: I don't share my backtest data on intraday candlestick patterns with anybody...so don't waste your time in asking for such. Backtesting is just a tool to give me some reference but doesn't tell me if I will make or lose money because I'm a discretionary trader. Also...there are intraday candlestick patterns that suck...those I don't trade) http://www.ttrader.com/mycharts/display_album.php?id=433 In the above link...EVERY Bearish Engulfing pattern that appeared since March 10th Monday...was profitable. I repeat...EVERY Bearish Engulfing pattern that appeared since March 10th Monday...was profitable. I traded some of them (not all). (I'll post tomorrow's chart if/when a Bearish Engulfing intraday candlestick pattern appears...regardless if I trade it or not...usually I can catch about 75% each trading week...they appear on average...once per day via my trade methodology.) Also, one of those trading days had 2 Bearish Engulfing pattern but I didn't post the chart because it appeared after 4pm est prior to the Eminis closing at 4:15pm est.... I always ignore candlestick trade signals in that 15min time window. Traderkay...once again...you can review my prior post about intraday candlesticks: http://www.elitetrader.com/vb/showthread.php?s=&postid=216063#post216063 By the way...after a quick research...I found that post where you made a similar comment about Bollinger Bands (I mistakenly made reference to it in this thread to imply it was about candlesticks): http://www.elitetrader.com/vb/showthread.php?s=&threadid=9997&perpage=1&pagenumber=7 Traderkay...here's your reply: http://www.elitetrader.com/vb/showthread.php?s=&threadid=9997&perpage=1&pagenumber=23 Thus, I've come to the conclusion that you really aren't interested in knowing how something works for someone else after it doesn't work for you... eventhough you have said in the past... "...No need for PM. I feel comfortable talking in public about this." Invitation is still open for you to disclose how you use Japanese Candlesticks... I and others will help if you have sincere questions. P.S. I'm not a scalper...Bearish Engulfing patterns often produce quick points (not ticks) as profits...best to cover and bank the money at first signs of trouble...and it's not my bread-n-butter intraday candlestick pattern. NihabaAshi
You said that my suggested way of backtesting is wrong. It's not wrong, and I'm not the one who came up with it. It's a widely accepted method to backtest any setup for an edge. First you define the entry with strict rules. (If you don't like using Overbought Stoch as indication of prolonged uptrend, you can use MACD or whatever else you feel is a good indicator of a trend). Then you simply backtest then entry with exits at fixed time intervals such as +1, +5, +10, +15, etc. bars. You look at win% at those intervals. If they're inferior, there's no edge and there's no sense in pursuing the setup further. Now I'm wondering how you backtested intraday candlesticks?