How does Sum function work

[Scotty123]

Hi,

I am trying to find stocks that are up 2 of the last 3 days, starting from yesterday. If you can help me do that, you can ignore my question below.

Thanks for any help!
Scott

Question Below:

I think that I can use the Sum function for my problem. But in my testing, I was surprised to see that stocks that were up for the last 3 days (yesterday, the day before yesterday and the day before that), had values of 1, 2 or 3 when I added a column with the code below.

def priceCondition = close[1] > close[2] and close[2] > close[3] and close[3] > close[4];
plot out = Sum(priceCondition, 3);

Can anyone explain why that is?

 

[BenTen]

I don't think you need to use the Sum function for this.

You can create your own pattern using the Candlestick Pattern Editor in TOS.

# Generation time: 2025-03-06T00:18:14.381729600Z

def IsUp = close > open;
def IsDown = close < open;
def IsDoji = IsDoji();
def avgRange = 0.05 * Average(high - low, 20);
plot PatternPlot =
    IsUp[2] and
    IsUp[1] and
    IsUp[0] and
    close[2] < close[1] and
    close[1] < close[0];

PatternPlot.SetPaintingStrategy(PaintingStrategy.BOOLEAN_WEDGE_UP);
PatternPlot.SetDefaultColor(GetColor(0));

 

[Scotty123]

I don't think you need to use the Sum function for this.

You can create your own pattern using the Candlestick Pattern Editor in TOS.

# Generation time: 2025-03-06T00:18:14.381729600Z

def IsUp = close > open;
def IsDown = close < open;
def IsDoji = IsDoji();
def avgRange = 0.05 * Average(high - low, 20);
plot PatternPlot =
    IsUp[2] and
    IsUp[1] and
    IsUp[0] and
    close[2] < close[1] and
    close[1] < close[0];

PatternPlot.SetPaintingStrategy(PaintingStrategy.BOOLEAN_WEDGE_UP);
PatternPlot.SetDefaultColor(GetColor(0));

View attachment 24239

Hi Ben,

Thanks for your quick reply.

I'm wanting code that will be used in a filter to return stocks that have gone up 2 of the past 3 days, starting yesterday.

I can't be sure, but this code seems to work.

def priceCondition = close[1] > close[2];
plot signal = Sum(priceCondition, 3) >= 2;

I believe it reads that the priceCondition is defined as yesterdays close is greater than the previous days close. And the Sum function says to filter the previous 3 days and output stocks that meet the priceCondition twice.

Does that seem right?

 

[halcyonguy]

I don't think you need to use the Sum function for this.

You can create your own pattern using the Candlestick Pattern Editor in TOS.

# Generation time: 2025-03-06T00:18:14.381729600Z

def IsUp = close > open;
def IsDown = close < open;
def IsDoji = IsDoji();
def avgRange = 0.05 * Average(high - low, 20);
plot PatternPlot =
    IsUp[2] and
    IsUp[1] and
    IsUp[0] and
    close[2] < close[1] and
    close[1] < close[0];

PatternPlot.SetPaintingStrategy(PaintingStrategy.BOOLEAN_WEDGE_UP);
PatternPlot.SetDefaultColor(GetColor(0));

the code doesn't match the image.
the code deal with open and close., starting with the current bar.
the image deals with comparing close, bar to bar, but still starts with current bar.
he wants to start with previous bar.

 

[halcyonguy]

Hi,

I am trying to find stocks that are up 2 of the last 3 days, starting from yesterday. If you can help me do that, you can ignore my question below.

Thanks for any help!
Scott

Question Below:

I think that I can use the Sum function for my problem. But in my testing, I was surprised to see that stocks that were up for the last 3 days (yesterday, the day before yesterday and the day before that), had values of 1, 2 or 3 when I added a column with the code below.

def priceCondition = close[1] > close[2] and close[2] > close[3] and close[3] > close[4];
plot out = Sum(priceCondition, 3);

Can anyone explain why that is?

because you are evaluating 5 comparisons, not 3

your formula looks at 3 comparisons.
then by using sum(), you are looking at 3 different sets of bars, offsets, 123, 234, 345
so 5 different comparisons

 

[halcyonguy]

Hi Ben,

Thanks for your quick reply.

I'm wanting code that will be used in a filter to return stocks that have gone up 2 of the past 3 days, starting yesterday.

I can't be sure, but this code seems to work.

def priceCondition = close[1] > close[2];
plot signal = Sum(priceCondition, 3) >= 2;

I believe it reads that the priceCondition is defined as yesterdays close is greater than the previous days close. And the Sum function says to filter the previous 3 days and output stocks that meet the priceCondition twice.

Does that seem right?

i agree with this .
but i like to make formulas that interact with current bar. then use an offset on the variable name , as needed. so my version of sum uses this , priceCondition[1], to start at the previous bar.

i made this as a lower, for testing
added a bubble to show count, to verify

declare lower;
#def priceCondition = close[1] > close[2];
#plot signal = Sum(priceCondition, 3) >= 2;

def priceCondition = close[0] > close[1];
plot signal = Sum(priceCondition[1], 3) >= 2;

addchartbubble(1,0, Sum(priceCondition[1], 3) , color.yellow, yes);

 

[Scotty123]

i agree with this .
but i like to make formulas that interact with current bar. then use an offset on the variable name , as needed. so my version of sum uses this , priceCondition[1], to start at the previous bar.

i made this as a lower, for testing
added a bubble to show count, to verify

declare lower;
#def priceCondition = close[1] > close[2];
#plot signal = Sum(priceCondition, 3) >= 2;

def priceCondition = close[0] > close[1];
plot signal = Sum(priceCondition[1], 3) >= 2;

addchartbubble(1,0, Sum(priceCondition[1], 3) , color.yellow, yes);

Thanks!