How to calculate RSI in SQL Server? - sql

Question

Can somebody give me an idea how to calculate the RSI (Relative Strength Index)?
The formula for RSI:
100 - 100 / (1 + RS)
RS (Relative Strength) = AverageGain / AverageLoss
RSI calculation is based on 14 periods. Losses are expressed as positive values, not negative values.
The very first calculations for average gain and average loss are simple 14-period averages:
First Average Gain = Sum of Gains over the past 14 periods / 14
First Average Loss = Sum of Losses over the past 14 periods / 14
The second, and subsequent, calculations are based on the prior averages and the current gain loss:
Average Gain = [(previous Average Gain) x 13 + current Gain] / 14.
Average Loss = [(previous Average Loss) x 13 + current Loss] / 14.
Below is my query which is not working because something in my recursion part is wrong. The way I calculate ROW_NUMBER: my first row is relative to the most recent trading day. I also make calculations over a 250 periods because of additional explanations, not relative to the calculation:
WITH bas AS
(
SELECT *
FROM
(SELECT
p.CurrentTicker, p.ClosePrice, p.PriceDate, p.ClosePriceChange,
(CASE WHEN p.ClosePriceChange > 0 THEN p.ClosePriceChange ELSE 0 END) Gain,
(CASE WHEN p.ClosePriceChange < 0 THEN ABS(p.ClosePriceChange) ELSE 0 END) Loss,
rownum
FROM
(SELECT
*,
ROW_NUMBER() OVER (PARTITION BY CurrentTicker ORDER BY PriceDate DESC) rownum
FROM
PriceHistory
WHERE
PriceDate >= DATEADD(day, -400, GETDATE())) p
INNER JOIN
Ticker t ON p.CurrentTicker = t.CurrentTicker
WHERE
rownum <= 250) a
),
rec AS
(
SELECT
CurrentTicker, PriceDate, ClosePriceChange, rownum,
AvgGain = SUM(Gain) OVER (PARTITION BY CurrentTicker ORDER BY PriceDate ROWS BETWEEN 13 PRECEDING and CURRENT ROW) / 14,
AvgLoss = SUM(Loss) OVER (PARTITION BY CurrentTicker ORDER BY PriceDate ROWS BETWEEN 13 PRECEDING and CURRENT ROW) / 14
FROM
bas
WHERE
rownum = (SELECT MAX(rownum) - 14 FROM bas)
UNION ALL
SELECT
bas.CurrentTicker, bas.PriceDate, bas.ClosePriceChange, bas.rownum,
a.AvgGain, a.AvgLoss
FROM
rec
INNER JOIN
bas ON rec.rownum - 1 = bas.rownum
AND rec.CurrentTicker = bas.CurrentTicker
CROSS APPLY
(SELECT
(rec.AvgGain * 13 + bas.Gain) / 14 AS AvgGain,
(rec.AvgLoss * 13 + bas.Loss)/14 as AvgLoss) a
)
SELECT *
FROM rec
WHERE CurrentTicker = 'AAPL'
ORDER BY PriceDate DESC
OPTION (MAXRECURSION 0)
Here is AAPL prices table:
PriceDate
ClosePrice
2021-02-23
125.86
2021-02-22
126
2021-02-19
129.87
2021-02-17
130.84
2021-02-16
133.19
2021-02-12
135.37
2021-02-11
135.13
2021-02-10
135.39
2021-02-09
136.01
2021-02-08
136.91
2021-02-05
136.76
2021-02-04
137.39
2021-02-03
133.94
2021-02-02
134.99
2021-02-01
134.14
2021-01-29
131.96
2021-01-28
137.09
2021-01-27
142.06
2021-01-26
143.16
2021-01-25
142.92
2021-01-22
139.07
2021-01-21
136.87
2021-01-20
132.03
2021-01-19
127.83
2021-01-15
127.14
2021-01-14
128.91
2021-01-13
130.89
2021-01-12
128.8
2021-01-11
128.98
2021-01-08
132.05
2021-01-07
130.92
2021-01-06
126.6
2021-01-05
131.01
2021-01-04
129.41
2020-12-31
132.69
2020-12-30
133.72
2020-12-29
134.87
2020-12-28
136.69
2020-12-24
131.97
2020-12-23
130.96
2020-12-22
131.88
2020-12-21
128.23
2020-12-18
126.655
2020-12-17
128.7
2020-12-16
127.81
2020-12-15
127.88
2020-12-14
121.78
2020-12-11
122.41
2020-12-10
123.24
2020-12-09
121.78
Now I hope someone will be able to help me. Thanks in advance.