Candlestick and Price Action Observations

[sethmo]

I've been studying candlesticks for a while and because I'm quite lazy and find it cumbersome to actually memorise formations, I've made a few observations:

1. Bullish patterns will pierce the high or the open > close.
2. Bearish patterns will pierce the low and open < close.

If you must know any candles intimately, focus on the doji and the hammer/hanging man. Either way, one of the two observations above will apply when preparing to place a trade.

On an interesting side note, you can make R5000 into R1 000 000 in 8 months if you double your starting capital each month. For example, R5000 - R 10 000; R10 000 - R20 000; R20 000 - R40 000, etc. This can be done by achieving 25% per week (100% per month). This is excluding costs (just for the sake of simplicity). Whether this is realistic or not remains to be seen, but I think it's worth trying.

 

[tomorton]

I like candlesticks also and am wary of complex indicators. I look for 2-day candlestick formations that mark the first two closes beyond the 14EMA. Continuation will usually follow.

For Longs, look for first 2 consecutive closes above 14EMA and go long on C2 if at least 6 of the following points apply -
C1>O1
C1>(H1-L1)/2
C2>O2
C2>C1
C2>H1
C2>(H2-L2)/2
H2>H1
L2>L1

If the second close is not on the same side of the 14EMA as the first, the pattern is not completed, ignore. There is a max score of 8 of course, and patterns scoring 5 or less are rarely followed by continuation. Works for both longs and shorts, but reverse the symbols for shorts.

 

[timsk]

On an interesting side note, you can make R5000 into R1 000 000 in 8 months if you double your starting capital each month. For example, R5000 - R 10 000; R10 000 - R20 000; R20 000 - R40 000, etc. This can be done by achieving 25% per week (100% per month). This is excluding costs (just for the sake of simplicity). Whether this is realistic or not remains to be seen, but I think it's worth trying.

Hi sethmo,
Welcome to T2W,
Your fantasy (quoted) is possible on paper, but it's a rather different story in reality! Without wishing to dampen your excitement and enthusiasm, I suggest you think in terms of breaking even in your first year, equalling or bettering a basic bank or building society interest rate in the second year and then a multiple of this in the third - say two times the interest rate. Hopefully, you'll do a lot better than this, in which case you can congratulate yourself in the knowledge that your results are better than most!
Cheers,
Tim.

 

[sethmo]

@tommorton

I understand your strategy to a certain extent. What it seems like is a formation that I like using (which is basically a modification of 3 white soldiers). I understand all the other conditions but I don't understand how the C1>(H1-L1)/2 and C2>(H2-L2)/2 conditions. Surely the daily range divided by two will rarely exceed the close price.

@timsk

It's nice to dream, hehehe. It's not hard to beat the interest on savings accounts in my country (5% p.a.) and fixed deposits (just under 10% and about 11% if you're over 65 - and I'm not waiting another 30-odd years for that to happen!). So if I do better than 5% p.a., that's good, over 10%, that's really good.

 

[tomorton]

@tommorton

I understand your strategy to a certain extent. What it seems like is a formation that I like using (which is basically a modification of 3 white soldiers). I understand all the other conditions but I don't understand how the C1>(H1-L1)/2 and C2>(H2-L2)/2 conditions. Surely the daily range divided by two will rarely exceed the close price.

@timsk

It's nice to dream, hehehe. It's not hard to beat the interest on savings accounts in my country (5% p.a.) and fixed deposits (just under 10% and about 11% if you're over 65 - and I'm not waiting another 30-odd years for that to happen!). So if I do better than 5% p.a., that's good, over 10%, that's really good.

Hi sethmo - Quite right, this is effectively TWO white soldiers, but with the added modifier of 14EMA breach (to represent upward price action, and hopefully trend) with a second close above 14EMA as confirmation.

On C1>(H1-L1)/2 and C2>(H2-L2)/2 conditions, the idea is that, for a long, the respective close is above the mid-point of the day's range, this being a positive sign of buying pressure. Bear in mind that one or other candlestick might not be a white soldier -though that would be ideal it actually could be a down day or something like an inverted hammer. I was thinking this moring of drawing up a visual representation of the pattern with various points values - I think I should do this and will post it somewhere when I do.

I did a backtest of this pattern on the Dow this morning for the whole of 2009 and posted results on the Swingin' the FTSE thread. 24 signals over the year, but 12 had scores of less than 6 and / or repeated or countered running positions from earlier signals. Of the 12 signals that should have been ignored, 3 only would have been winners: of the 12 trades taken, 9 would have been winners. The net gain from the 12 confirmed trades was +2,500 Dow points. As you can imagine, I am very pleased with this result, having been searching a simple and virtually automatic chart signal for years now. On next to backtest on Spot Gold 2009, then other markets.

 

[timsk]

Hi Tom,
You're candle pattern looks straightforward and mechanical to implement - which is great. Can you please clarify the results on the test you've done? So far, what you've outlined to us is an entry technique. How did you define a failed pattern and a successful one respectively and how did you arrive at the impressive +2,500 Dow points? Presumably you applied some sort of exit strategy. If this is asking you to divulge 'trade secrets' which you'd rather not do - that's fine, I understand.
Cheers,
Tim.

 

[tomorton]

Hi Tom,
You're candle pattern looks straightforward and mechanical to implement - which is great. Can you please clarify the results on the test you've done? So far, what you've outlined to us is an entry technique. How did you define a failed pattern and a successful one respectively and how did you arrive at the impressive +2,500 Dow points? Presumably you applied some sort of exit strategy. If this is asking you to divulge 'trade secrets' which you'd rather not do - that's fine, I understand.
Cheers,
Tim.

Hi Tim -

To generate a trade, the pattern has to score 6 or more out of 8, as per the points outlined. A perfect 8-pointer might look like two white soldier candlesticks, the first of which breaches the 14EMA, closing above the average after a string of closes below. I will post some chart examples soon but meantime see Dow for 05 and 06/11/09. 05 gets 2/2 points - close above open and close above range's mid-point. 06 gets 6/6 points - close above open, close above close of 05, close above high of 05, close above mid-point of range of 06, high above high of 05, low above low of 05.

If 06 had closed back below the 14EMA, this would not be a signal at all. A signal scoring 5 or less would be ignored - e.g. the pattern of 08-09/12 scores 4 only. A signal that counters or repeats the signal for a running trade is also ignored.

Only two possible exit scenarios -
1) stop-loss is breached: for longs, this is low of day 1, for shorts high of day 1
or
2) first close which is in profit but which is back below 14EMA for longs, above 14EMA for shorts

Exit 2 is conservative. Taking the entry signal on 06/11 at 10023, the stop-loss (9808, low of 05/11) was not hit, price running up to make a close back below the 14EMA for the first time on 08/12 at 10286, making a gain of +263. However, price had reached a high of 10516 on 04/12, so the potential gain on the trade was actually +493. My 2009 net gain of +2,561 is based on the conservative closing price level, and is way below the potential maximum gain.

Actually, the 2009 Dow chart shows 24 signals. 12 of these are either weak, scoring 5 pattern points or less or they counter or repeat signals on running trades, and so would have been ignored. However, taking all 24 trades would still have given a profit of +1,860, not that I recommend this, but I take this to show the inherent strength of this simple pattern.

 

[timsk]

Cheerrs Tom - good work!
Tim.

 

[tomorton]

2009 review for this method on Spot Gold is ambiguous.
6/14 confirmed trades were winners, for horrible net loss -$184.50.
But taking the weak, counter and repeat signals gave 10/26 signals over the year, making net loss just -$18.35.

I suppose that, like other trend-following systems, this method isn't that good when markets are not trending.