StockFetcher Forums · General Discussion · True Strength Index (TSI)<< >>Post Follow-up
glgene
535 posts
msg #96468
Ignore glgene
9/23/2010 1:23:06 AM

Note: This thread originally began as "PPO vs. MACD" in the General Discussion Forum
--------------------------------------------------------------------------
OK, Kevin_in_GA. I read TSI (True Strength Index) info on page 107 in the Stockfetcher Manual, and I also Googled TSI to read more info on it. I read where the TSI values of -25 to +25 correspond to typical RSI readings of 30 and 70 for "oversold" and "overbought" levels. You had mentioned the TSI values range from -100 to +100. Is that range from SF?

Anyway, here is the script I came up with (with beginning help from you, of course). It may appear a little clunky with my added lines, but it's the best I can do at the moment. I welcome constructive input.

Fetcher[
symlist(vti)
draw TSI(12,26,9)
draw MACD(12,26,9)
add column MACD(12,26,9)
add column TSI(12,26,9)
add column smoothed TSI(12,26,9)
set{t,TSI(12,26,9)}
set{s,smoothed TSI(12,26,9)}
set{r, t-s}
add column r{tsi-Smoothed tsi}
]



Some notes:
--------------------
1) I originated a 'set' variable (r) that shows the numeric difference between the TSI and Smoothed TSI values, so that can be shown in a column. That column heading is "tsi-smoothed tsi." Going from above 0 to below 0 would be a selling crossover, and vice versa. I need that column info, so I can add it to my daily Tracker info I maintain in a spreadsheet, along with other items I track.

2) Your draw command of TSI displays the info quite nicely. Thank you! How would you add to the script to chart display, via a 'draw' command, the 'Difference' number I show in the tabled info?

3) Here is some 2010 historical info on VTI (ETF for Vanguard Total Stock Index), using the above script:

.......................MACD .............TSI ...... Smoothed TSI ........ Diff.

Aug. 10 .......... 0.59 .......... 15.84 .......... 12.03 ................ +3.81

Aug. 11 .......... 0.44 .......... 11.60 .......... 11.94 ................. -0.34

Aug. 17 ......... 0.0059 ....... 0.16 .............. 6.35 ................. -6.19

Aug. 18 ......... -0.0056 ..... -0.15 ............. 5.05 .................. -5.20

Sept. 1 ........... -0.50 ......... -14.22 ........ -12.15 ................ -2.07

Sept. 2 ........... -0.37 ......... -10.16 ........ -11.75................ +1.60

Sept. 22 ........ 0.77 ........... 23.61 .......... 15.21 ................ +8.40

4) TSI (12,26) didn't go negative until Aug. 18 (-0.15), but TSI -Smoothed TSI (12,26,9) established a negative difference on Aug. 11 (-0.34 diff.). Which of those dates, Kevin, would trigger a sale (if you were using this indicator by itself); Yes, I suspect you would use confirming indicators, but that's not the question here. At this point, with my limited work with TSI, I favor using the 'difference' column as the trigger point (taking advantage of the 9-period smoothing factor.

5) On Sept. 22, the TSI number is 23.61. If the article I read that says the top range number is +25, then that would appear we're hitting the 'overbought' ceiling very soon. But if the real ceiling is +100, as you said, then we've got a long way to go. Comments?

6) An even more basic question for you: Did I calculate the Smoothed TSI number correctly? I think I did, but I would like you to confirm (or correct me).

7) The above script shows TSI with 12,26,9 (typical MACD) settings. I have read where the normal TSI setting is 25,13, 9 (or is there a better number than 9 as the smoothing component?). What settings do you use, if you care to share?

8) Again, if you care to share it, how do you use TSI (True Strength Index) in your day-to-day stock analysis? Or anyone else here?

Thank you for your beginning work with me on this subject (PPO vs. MACD), and thank you in advance for your continued work here. You're a real gem, as so many posters here are. I envy all of your scripting expertise. I'm still at the 101 level.

Zack

mystiq
650 posts
msg #96471
Ignore mystiq
9/23/2010 2:09:54 AM

8) Again, if you care to share it, how do you use TSI (True Strength Index) in your day-to-day stock analysis? Or anyone else here? ---> since you asked: i like to use TSI crossing ZERO for the trigger

glgene
535 posts
msg #96472
Ignore glgene
9/23/2010 2:23:39 AM

Mystiq,
TSI(12,26) crossing 0, or TSI (12,26) crossing smoothed TSI(12,26,9)?

If 12,26 aren't your settings, do you mind saying what they are?

Zack

Kevin_in_GA
4,553 posts
msg #96479
Ignore Kevin_in_GA
9/23/2010 11:17:28 AM

OK, Kevin_in_GA. I read TSI (True Strength Index) info on page 107 in the Stockfetcher Manual, and I also Googled TSI to read more info on it. I read where the TSI values of -25 to +25 correspond to typical RSI readings of 30 and 70 for "oversold" and "overbought" levels. You had mentioned the TSI values range from -100 to +100. Is that range from SF?

If you look at how the TSI is constructed, it is bounded at 100 and -100 by definition. The suggested OS/OB values are just guides. I would use the crossover between the TSI and its EMA(9) - what you are calling the "smoothed" TSI.

Anyway, here is the script I came up with (with beginning help from you, of course). It may appear a little clunky with my added lines, but it's the best I can do at the moment. I welcome constructive input.


Fetcher[
symlist(vti)
draw TSI(12,26,9)
draw MACD(12,26,9)
add column MACD(12,26,9)
add column TSI(12,26,9)
add column TSI histogram(12,26,9)
]



I modified your filter just a bit to make it cleaner - you can simply use "TSI Histogram" instead of the set{} commands you were using. The new column (automatically called diff) is the difference between the TSI and its smoothed variant.

Some notes:
--------------------
1) I originated a 'set' variable (r) that shows the numeric difference between the TSI and Smoothed TSI values, so that can be shown in a column. That column heading is "tsi-smoothed tsi." Going from above 0 to below 0 would be a selling crossover, and vice versa. I need that column info, so I can add it to my daily Tracker info I maintain in a spreadsheet, along with other items I track.

2) Your draw command of TSI displays the info quite nicely. Thank you! How would you add to the script to chart display, via a 'draw' command, the 'Difference' number I show in the tabled info?

draw TSI Histogram(12,26,9) line at 0

3) Here is some 2010 historical info on VTI (ETF for Vanguard Total Stock Index), using the above script:

At this point, with my limited work with TSI, I favor using the 'difference' column as the trigger point (taking advantage of the 9-period smoothing factor.

Me too. Just use the crossover between the TSI "fast line" and "slow line" like you would with the MACD.

5) On Sept. 22, the TSI number is 23.61. If the article I read that says the top range number is +25, then that would appear we're hitting the 'overbought' ceiling very soon. But if the real ceiling is +100, as you said, then we've got a long way to go. Comments?

Those values are arbitrary, and depend on the input settings. I would play around with input numbers, and once I found one that crossed its slow line at the right times, just use that instead.

6) An even more basic question for you: Did I calculate the Smoothed TSI number correctly? I think I did, but I would like you to confirm (or correct me).

Yes, but referencing the TSI Histogram is easier.

7)What settings do you use, if you care to share?

I use this indicator to tell me the relative strength of a series of ETFs that act as proxies for different asset classes (smallcaps, largecaps, emerging markets, bonds, gold). I use weekly charts for this, and I have backtested this indicator to find the best settings for my particular sytle of investing. The approach is fully laid out in the 401k thread I posted here, but it simply says to buy the top scoring ETF based on its highest TSI score.

8) Again, if you care to share it, how do you use TSI (True Strength Index) in your day-to-day stock analysis? Or anyone else here?

Weekly TSI(5,5,9). These settings are excellent for catching trends. While no one indicator will always work, this is one in which I have real confidence.

Thank you for your beginning work with me on this subject (PPO vs. MACD), and thank you in advance for your continued work here. You're a real gem, as so many posters here are. I envy all of your scripting expertise. I'm still at the 101 level.


glgene
535 posts
msg #96491
Ignore glgene
9/23/2010 5:51:50 PM

Kevin,
Thanks for your prompt and thorough follow-up to my questions.

Follow-up from me:

1) Where, within TSI, does it say it's range-bound between -100 and +100. Sorry for my lack of understanding.

2) I discovered that when you use this command (shown below) it produces the same results.

-------- Add column TSI (12,26)
-------- Add column TSI (12,26,9)

But the Diff is correct.

3) Same with MACD (12,26) and MACD (12,26,9). Same column results. ????

To help prove that point, I added these 2 columns to the script below:

EMA 12
EMA 26

Subtract the above two and you get MACD(12,26).

(I added SPY, DIA and QQQ to the original Symlist VTI)

Fetcher[
symlist(vti,spy,dia,qqqq)
draw TSI(12,26,9)
draw MACD(12,26,9)
add column ema(12)
add column ema(26)
add column MACD(12,26)
add column MACD(12,26,9){macd 12,26,9}
add column TSI(12,26)
add column TSI(12,26,9){tsi 12,26,9}
add column TSI histogram(12,26,9)
]



4) If you use Weekly TSI(5,5,9), how do the two identical 5s relate to each other in the script?

5) When you view the MACD and TSI charts in alignment, both at 12-26-9, they certainly look identical. If so, what is the value of using TSI when MACD produces the same trade date results, other than with TSI you can compare different stocks in the tabled view in an apples-to-apples comparison (which was my first endeavor with finding something to replace MACD)? What is the main value of TSI from your perspective?

As usual, a hearty thanks to you for helping educate someone at the 101 level (me). I still haven't given up on the idea of modifying the PPO script to get identical trade dates with MACD (but first, TSI). How did you become so proficient with SF?

Zack




Kevin_in_GA
4,553 posts
msg #96492
Ignore Kevin_in_GA
modified
9/23/2010 7:46:13 PM

1) Where, within TSI, does it say it's range-bound between -100 and +100. Sorry for my lack of understanding.

Look at how the indicator is constructed - "The TSI first creates a double EMA of the difference between the daily closing values. This value is divided by the double-smoothed absolute difference of the daily closing values. The raw signal is accompanied by a smoothed version and a histogram illustrating the difference between the two values." Dividing by the absolute difference guarantees bounds at 100 and -100. Think about it.

2) I discovered that when you use this command (shown below) it produces the same results.

-------- Add column TSI (12,26)
-------- Add column TSI (12,26,9)

But the Diff is correct.

Yes. That is why I suggested using the "TSI Histogram" command - it gives you the difference you were looking for. MACD Histogram works the same way.

4) If you use Weekly TSI(5,5,9), how do the two identical 5s relate to each other in the script?

Look at how the indicator is constructed - you'll answer this question on your own easily by reviewing the equations.

5) When you view the MACD and TSI charts in alignment, both at 12-26-9, they certainly look identical. If so, what is the value of using TSI when MACD produces the same trade date results, other than with TSI you can compare different stocks in the tabled view in an apples-to-apples comparison (which was my first endeavor with finding something to replace MACD)? What is the main value of TSI from your perspective?

The TSI is a measure of intrinsic strength - I looked at it as a possible replacement for the Comparative Relative Strength command, which compares the stock or ETF to another stock (e.g., the SPY). The RS indicator is normalized against a benchmark, but if the stock is falling but not as fast as the SPY, it looks like it is doing well rather than poorly.

As usual, a hearty thanks to you for helping educate someone at the 101 level (me). I still haven't given up on the idea of modifying the PPO script to get identical trade dates with MACD (but first, TSI). How did you become so proficient with SF?

I read the manual, posted questions only when I could not find the answers myself, and tried to help others write filters (which made me learn a lot of the SF coding tricks). You learn a lot doing these things, and can give back to those who helped you when you first started out.


mystiq
650 posts
msg #96501
Ignore mystiq
9/24/2010 1:51:04 AM

glgene
msg #96472
- Ignore glgene 9/23/2010 2:23:39 AM

Mystiq,
TSI(12,26) crossing 0, or TSI (12,26) crossing smoothed TSI(12,26,9)?

If 12,26 aren't your settings, do you mind saying what they are?

Zack

TSI(2,3,1) crossing zero

glgene
535 posts
msg #96538
Ignore glgene
9/24/2010 2:26:10 PM

Mystiq,
You must do a LOT of trading with your TSI(2,3,1) setting.

???

mystiq
650 posts
msg #96556
Ignore mystiq
9/25/2010 12:51:16 AM

glgene
msg #96538
- Ignore glgene 9/24/2010 2:26:10 PM

Mystiq,
You must do a LOT of trading with your TSI(2,3,1) setting.

???

not really - only used with other confirming indicators- really helpfull in judging the strength of the possible move-similar to a kickstart on a bike.

StockFetcher Forums · General Discussion · True Strength Index (TSI)<< >>Post Follow-up

*** Disclaimer *** StockFetcher.com does not endorse or suggest any of the securities which are returned in any of the searches or filters. They are provided purely for informational and research purposes. StockFetcher.com does not recommend particular securities. StockFetcher.com, Vestyl Software, L.L.C. and involved content providers shall not be liable for any errors or delays in the content, or for any actions taken based on the content.


Copyright 2016 - Vestyl Software L.L.C.Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus