-
The Forum works well on MOBILE devices without an app: Just go to: https://forum.tsptalk.com
-
Please read our AutoTracker policy on the IFT deadline and remaining active. Thanks!
div>
$ - Premium Service Content (Info) | AutoTracker Monthly Winners | Is Gmail et al, Blocking Our emails?
Find us on: Facebook & X | Posting Copyrighted Material
Join the TSP Talk AutoTracker: How to Get Started | Login | Main AutoTracker Page
The Forum works well on MOBILE devices without an app: Just go to: https://forum.tsptalk.com ...
Or you can now use TapaTalk again!S-fund Challenge! Guesses
- Thread starter ebbnflow
- Start date
userque
TSP Legend
- Reaction score
- 36
Last edited:
userque
TSP Legend
- Reaction score
- 36
Re: S-fund Challenge!
Another late night guess change: UP 1007.28 (Hopefully, I won't regret this change as must as I did a previous change I made :worried
I love it when the NN's differ with the EbbCharts
(Once I finish this latest updating/configuring of the NN's, they will be running continuously without being 'reset,' as they are being now. This will allow them to save and build from previous learning, rather than begin started anew each day or less. Then they should become 'smarter' and smarter [by a smaller and smaller factor] each day; and be much less likely to "change their minds." That's the hope at least...the challenge will reveal.Last edited:
userque
TSP Legend
- Reaction score
- 36
Re: S-fund Challenge!
Last guess: 1008.58
Just overhauled the NN, so it likely won't be able to beat the 'accuracy'(/probability/reliability) of it's above guess before market open.
UPDATE: Correction: The S-Fund Challenge NN hasn't been overhauled. I overhauled the F-Fund NN. S-Fund Challenge NN still has all of its wits and may still give a better guess before long.Last edited:
ebbnflow
Ebbchart
- Reaction score
- 30
Re: S-fund Challenge!
Another late night guess change: UP 1007.28 (Hopefully, I won't regret this change as must as I did a previous change I made :worried
I love it when the NN's differ with the EbbCharts.
(Once I finish this latest updating/configuring of the NN's, they will be running continuously without being 'reset,' as they are being now. This will allow them to save and build from previous learning, rather than begin started anew each day or less. Then they should become 'smarter' and smarter [by a smaller and smaller factor] each day; and be much less likely to "change their minds." That's the hope at least...the challenge will reveal.
My feelings, exactly!
ebbnflow
Ebbchart
- Reaction score
- 30
Re: S-fund Challenge!
For Apr. 10, 2014:
S&P 500 gauge: 54.9% (7-yr. win percentage).
Pattern: 2/grn-grn-red (WP: C 55.1%, S 51.9%, I 53.4%).
Strategy: long > 54.9%; short <= 53%; cash > 53% and <= 54.9%.
Forecast: long C-fund; short S-fund; cash I-fund.
Worst pattern for the S-fund.
I did say it was the worst pattern for the S-fund.
For Apr. 11, 2014:
S&P 500 gauge: 54.9% (7-yr. win percentage).
Pattern: 6/red-red-grn (WP: C 52.7%, S 54.3%, I 53.7%).
Strategy: short <= 53%; cash > 53% and <= 54.9%; long > 54.9%.
Forecast: short C-fund; cash S-fund; cash I-fund.
userque
TSP Legend
- Reaction score
- 36
Last edited:
RealMoneyIssues
TSP Legend
- Reaction score
- 100
Re: S-fund Challenge!
Up
Sent from my Samsung Galaxy S4 using tapatalk...
userque
TSP Legend
- Reaction score
- 36
ebbnflow
Ebbchart
- Reaction score
- 30
Re: S-fund Challenge!
What was I thinking? When I began this challenge, my strategy was to go short if a pattern's win percentage is less than 54.9% (7-yr. S&P 500 win percentage) and go long if it's over 54.9%. The problem with this is I'll be shorting with 5 patterns and going long with only 3. And to go short on any pattern with a win percentage of over 53% is just wrong. Anything over 53% is bullish in my book, so it was a losing proposition.
Old strategy: short < 54.9%; cash = 54.9%; long > 54.9%.
New strategy: short <= 53%; cash > 53% and <= 54.9%; long > 54.9%.
Using the new strategy, my winning percentage would be 5/7 instead of 5/10. Yep, 3 of those guesses would have been in cash. The trading setup will have 3 short, 2 cash, and 3 long patterns.
Short patterns: 2/grn-grn-red (51.7%); 7/red-grn-grn (52.9%); 8/red-grn-red (52.4%).
Cash patterns: 4/grn-red-grn (53.8%); 6/red-red-grn (54.3%).
Long patterns: 1/grn-grn-grn (55.5%); 3/grn-red-red (56.7%); 5/red-red-red (60.1%).
userque
TSP Legend
- Reaction score
- 36
Re: S-fund Challenge!
The benefits of testing a system with a virtual trading platform rather than with just percentage calculations.
ebbnflow
Ebbchart
- Reaction score
- 30
Re: S-fund Challenge!
Hear, hear! The platform is open to anyone, so it's time to put your money where your mouth is.
ebbnflow
Ebbchart
- Reaction score
- 30
Re: S-fund Challenge!
I did say it was the worst pattern for the S-fund.
For Apr. 11, 2014:
S&P 500 gauge: 54.9% (7-yr. win percentage).
Pattern: 6/red-red-grn (WP: C 52.7%, S 54.3%, I 53.7%).
Strategy: short <= 53%; cash > 53% and <= 54.9%; long > 54.9%.
Forecast: short C-fund; cash S-fund; cash I-fund.
For Apr. 14, 2014:
S&P 500: 54.9% (7-yr. win percentage).
Pattern: 7/red-grn-grn (WP: C 53.9%, S 52.9%, I 48.5%).
Strategy: short <= 53%; cash > 53% and <= 54.9%; long > 54.9%.
Forecast: cash C-fund; short S-fund; short I-fund.
RealMoneyIssues
TSP Legend
- Reaction score
- 100
userque
TSP Legend
- Reaction score
- 36
userque
TSP Legend
- Reaction score
- 36
Re: S-fund Challenge!
After much tinkering under the hood; looks like DOWN to 960.25 for Monday.
userque
TSP Legend
- Reaction score
- 36
Re: S-fund Challenge!
While investigating (for hours) the recent performance of the NN, I discovered a major FUBAR! Some type of way, I shifted the spreadsheet cells in the all-important price output column. This is the sheet that is imported into the NN. Suppose I hadn't had enough to drink that day. Anyway, hopefully the madness is over and I can get back to wwwwinning!:toung:
NN 2.0 is currently crunching, will post results later.
UPDATE:
That was fast. When the input makes sense, NN can figure things out nicely. While continued crunching will likely change the target, it's UP forecast, I'm suspecting, will remain...we'll see...maybe, as it's not projecting to be up by that much right now.
UP, target: 972.22Similar threads
- Article
