Seabridge Gold April 6, 2015

alcerec

What would the proper scan to capture the action of Seabridge Gold stock today (April 6,2015) with the prior day doji then a strong gap up?

markd

Apparently, the scan engine does not recognize the 4/2 bar for SA as a doji, but I don't know why.

If you write:

[group is GoldMining]
and [1 day ago doji is true]

it should pick up SA, but it doesn't. It does pick up other hits with apparently similar bars.

I tested

and [1 day ago doji is true]

with other groups and it does pick up one day ago dojis, so that phrase seems to be valid.

So, if the 4/2 bar were recognized as a doji, the scan would be:

[group is GoldMining]
and [1 day ago doji is true]
and [low > 1 day ago high]

The [group is GoldMining] is necessary because the next two conditions alone return more than 999 hits. You could use instead

[group is MaterialsSector]
and [close > 3]
and [volume > 40000]
... etc.

Maybe you could ask support why that 4/2 bar isn't a doji.

alcerec

thank you again for the help MarkD this really threw me. Could you explain the 4/2 a little further.

markd

4/2 means April 2, the date of the bar that looks like it should be a doji.

gord

I'm assuming the open and close are not close enough together to qualify as a doji. SA is about 0.2%, when I scan for doji's I get stocks in the 0.1% range.

"Doji form when a security's open and close are virtually equal."
http://stockcharts.com/school/doku.php?id=chart_school:chart_analysis:candlestick_pattern_dictionary

markd

I thought that might be it, too, but Open =5.75, Last = 5.74. You would think that's close enough.

gord

Yes at this level 1 cent seems close, but on a percentage basis its alot if the stock is $0.10 and its small if the stock is $100.

markd

I understand the argument. Maybe SC should document it's scanning criteria for candle types, like [doji is true] when... so users like @alcerec (and me) can figure out why their scans aren't working when they appear to be totally correct.

gord

Yes Markd I agree it would be nice to have more documentation, in particular on the coding of candlestick scans, I'm sure its somewhere on the wish list. However I have to say that the documentation on StockCharts has increased tremendously in the last few years and you will never make everyone happy.

There have been a ton of new additions and features recently, many don't really need any documentation and some probably should have a little documentation. But I don't want to stop the introduction of new features just waiting for the detailed docs to be written.

The other problem is the more detailed your documentation is the more horsepower it takes to keep it all up to date, especially every time something changes on the site. One new feature or change could affect 10"s or 100's of documents.

I'll vote for just being able to see the candlestick code, like we can if we insert some of the other predefined scans, I think this would be acceptable to 99% of the interested users.

Of course in the end if we are really concerned about the exact scan, we can always write our own code, with our own settings.

Kevo

Weird and interesting...

1. I could not find any documentation that suggested the Doji calculation used Percent Change of the Open and Close.

2. I thought of testing what gord was saying because I did not understand what the "SA is about 0.2%" meant.

I ran this scan:

[type = stock] AND [Daily SMA(20,Daily Volume) > 40000]

and [Close < 10.00]
and [Close > 1.00]

and [Open is not Close]
and [Doji is true]


and got symbol PDLI with Open = 6.97 and Close = 6.96

Percent Change = (6.97 - 6.96)/6.97 = 0.14% = 0.1%, which is less than "0.2%".

5.75 and 5.74 are relatively close to 6.97 and 6.96, but it does not meet the "0.1%"; Fluke maybe. I guess I understand now.

Percent Change = (5.75 - 5.74)/5.75 = 0.17% = 0.2%

alcerec

you guys are truly guru's thanks for the discussion.

markd

@gord, agree with your diagnosis. Scarce resources means you can't do everything all at once. One of the real strengths of SC, and one that makes it unique is the quantity and quality of documentation that does exist.

Kevo

Tom Bowley is now with SCC full-time. It looks like he will work on ChartSchool:

Looking Good, Feeling Good! Tom Bowley Joins StockCharts.com