Ebb's Account Talk

[ebbnflow]

Pattern 1/grn-grn-grn knows best.

Thursday: Pattern 1/grn-grn-grn. Win Percentage (CSI 52.9%): C 53.4%, S 56.3%, I 49.0%.
Result: cash C-fund was safe; long S-fund was right; short I-fund was wrong.

Unofficial (May 29, 2014): C +0.54%; S +0.45%; I +0.29%.

Short Patterns (S-fund): 2/grn-grn-red (51.9%); 8/red-grn-red (52.3%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%); 7/red-grn-grn (53.1%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.1% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.1%; long > 55.1%.
Accuracy: 22/33 (67%). Long: 16/22 (73%). Short: 6/11 (55%). Cash: green (3 up, 3 down); red (3 up, 3 down).

Friday: Pattern 8/red-grn-red. Win Percentage (CSI 54.9%): C 55.6%, S 52.3%, I 56.8%.
Forecast: long C-fund; short S-fund; long I-fund.

Friday has pattern 8/red-grn-red. Win Percentage (S-fund): 52.3% (W 123, L 112). Strategy: Short <= 53%.

Today's market action has pattern 8/red-grn-red written all over it. It was a perfect snapshot (Kodak moment). :nuts:

Friday: Pattern 8/red-grn-red. Win Percentage (CSI 54.9%): C 55.6%, S 52.3%, I 56.8%.
Result: long C-fund was right; short S-fund was right; long I-fund was right.

Unofficial (May 30, 2014): C +0.18%; S -0.30%; I +0.05%.

Short Patterns (S-fund): 2/grn-grn-red (51.9%); 8/red-grn-red (52.1%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%); 7/red-grn-grn (53.1%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.1% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.1%; long > 55.1%.
Accuracy: 23/34 (68%). Long: 16/22 (73%). Short: 7/12 (58%). Cash: green (3 up, 3 down); red (3 up, 3 down).

Monday: Pattern 2/grn-grn-red. Win Percentage (CSI 53.3%): C 55.1%, S 51.9%, I 53.0%.
Forecast: long C-fund; short S-fund; short I-fund.

Monday has pattern 2/grn-grn-red. Win Percentage (S-fund): 51.9% (W 123, L 114). Strategy: Short <= 53%.

 

[userque]

Userque, 53% works for the patterns. Here's a link:

Winning Percentages

Although there are, indeed, propositions that offer more than a 60% expectation of winning, such propositions are relatively few and far between, and are only a very small part of the overall picture. With the break-even point at about 53%, genuine professional bettors know there is no tenable excuse to pass up propositions offering expectations of higher than, say, 55 percent. A small advantage applied over and over is awesomely effective. Mathematicians will confirm that a profit is more assured from a group of 200 bets with a 55% expectation-per-bet than from a group of 50 bets with a 60% expectation-per-bet. In other words, the more bets placed, the more predictable the outcome. (See our article, Binomial Distribution.)

Thanks, will check this out and get back to you about it.

 

[userque]

Userque, 53% works for the patterns. Here's a link:

Winning Percentages

Although there are, indeed, propositions that offer more than a 60% expectation of winning, such propositions are relatively few and far between, and are only a very small part of the overall picture. With the break-even point at about 53%, genuine professional bettors know there is no tenable excuse to pass up propositions offering expectations of higher than, say, 55 percent. A small advantage applied over and over is awesomely effective. Mathematicians will confirm that a profit is more assured from a group of 200 bets with a 55% expectation-per-bet than from a group of 50 bets with a 60% expectation-per-bet. In other words, the more bets placed, the more predictable the outcome. (See our article, Binomial Distribution.)

I've read the article.

If 53% worked for the patterns, wouldn't you simply say: ">=53% Long, <53% Short?"...all the time? Why go to cash if 53 is the magic number? According to the rule-of-53, you're leaving money on the table.

That 53% is purported to be based upon a set risk/reward: 11 to win 10. (Assuming it even works for that. No mathematically robust evidence is given in the article.) In contrast, the market, on any given day, has a varying risk/reward. It'll move up/down a certain varying percentage. Also, the amount of cash you will gain/lose depends upon your investment. It's not so clean with the markets.

To say this 53% number is universal is to say that all bets are 11 to win 10. Which of course, they aren't (lottery ticket anyone).

If you play/understand poker:

A tight player will have a lower winning percentage over time than a loose player (folding more hands); but will be more profitable over time. We've seen this effect in the Challenge here.

I don't think that 53% number is a constant that applies all over the universe.

The universal 53% theory is flawed, but rather than present a mathematical proof, I'll conduct an experiment. Here's what I'll do (another thing on the todo list). I'll build a sheet to simulate random trades on the W4500 since 2007 at a 53% success rate and see how it does. I can simulate all those trades with the click of a mouse. I will do several passes (clicks). I will even be able to change the percent after the initial test is complete...a what-if type of thing. This is the kind of optimization I was suggesting you do to obtain and maintain the best "numbers" for your system. You will be able to download the file if you like and run it for yourself.

After I debunk this myth (or maybe I'll prove it!?[SUP]1[/SUP]), maybe I'll look into the global warming thing then go see what they're talking about in that forum:blink:.....j/k


-------------------
[SUP]1[/SUP]The assertion is that 53% applies generally. Finding another example where is does apply (let's assume: the stock market), doesn't suggest that the original hypothesis is proven. However, finding just one example of it not applying, shows the hypothesis is incorrect. In simpler terms: I can't prove this 53% thingy correct; but I can debunk it.

 

[ebbnflow]

If I just go long on any pattern above 53%, then I wouldn't have a great record of 16/22 (73%) calling longs. I don't mind leaving some profits on the table by being in cash. Record for cash positions: green (3 up, 3 down); red (3 up, 3 down). So with cash/neutral patterns, I'm only hitting 50/50. Even my short patterns are doing better at 7/12 (58%).

Like they say, "if it ain't broke, don't fix it." But still, it would be nice to find the sweet spot for these patterns.

 

[userque]

If I just go long on any pattern above 53%, then I wouldn't have a great record of 16/22 (73%) calling longs. I don't mind leaving some profits on the table by being in cash. Record for cash positions: green (3 up, 3 down); red (3 up, 3 down). So with cash/neutral patterns, I'm only hitting 50/50. Even my short patterns are doing better at 7/12 (58%).

Like they say, "if it ain't broke, don't fix it." But still, it would be nice to find the sweet spot for these patterns.

Definitely can't argue with success.

 

[userque]

Re: Still Clueless?

Ebb,

I've updated the infrastructure of the tally spreadsheet to make future upgrades quicker and easier, should there be any. I'm also finishing up implementing the trading fees. Looks like $10,000 is still 'good enough.' You took a smaller than (I) expected hit: you're now around $10,400. Not bad at all.

My thoughts on the fees are:

$5 per trade. I assume we're trading retirement accounts; so generally, no shorting. However, we can trade inverse funds. So, from CASH to LONG/SHORT and vice versa, is one trade. From LONG to SHORT and vice versa would be two trades: exit one ETF, enter the inverse/opposite ETF...$10.

Let me know if you (or any other participant, btw) want me to post your detailed account history for your records/review. I don't use a separate column for the fees, they are incorporated into the balance formula. Basically, for each row/day, the fees are taken out last, as though a trade was made at closing.

The heavy lifting is now done. Next, I'll work on what we talked about a while back: letting the CASH position earn a little interest. That'll be easier if I just use a set daily interest rate rather than tie it to the G-Fund on a daily basis.

All of the above are/will-be retroactive.

Another idea I have is to use three times the percentage change of the S-Fund in calculating our trade profits/losses to simulate the real world, balls out, trading of 3X funds. This ain't yo grandpa's retirement fund.

Your thoughts?

UPDATE: Re: 3X funds. I guess I'm thinking the challenge ought not be just to beat the S-Fund, but to crush it. To make as much money as possible...or get felted trying. If you want safe, open a CD.

 

[ebbnflow]

Re: Still Clueless?

I'm all in with the 3x leveraged funds. That's what everybody does to offset fees. Balls to the wall, userque!

 

[userque]

Re: Still Clueless?

I'm all in with the 3x leveraged funds. That's what everybody does to offset fees. Balls to the wall, userque!

That change put you at 11.5K

I'll post the sheet later.

 

[ebbnflow]

Re: Still Clueless?

That change put you at 11.5K

I'll post the sheet later.

I wonder if trading less could be better (I'd like to find out). A strategy that follows the double, triple, quadruple patterns and not so much the daily signals (less fees). I'll call it the DTQ ByPass. Userque, can you include this trader/strategy to start this Monday (short, of course). Thanks.

 

[userque]

Re: Still Clueless?

I wonder if trading less could be better (I'd like to find out). A strategy that follows the double, triple, quadruple patterns and not so much the daily signals (less fees). I'll call it the DTQ ByPass. Userque, can you include this trader/strategy to start this Monday (short, of course). Thanks.

WILCO

 

[userque]

Re: Still Clueless?

Ebb,

FWIW

I ran the experiment. The experiment simulates randomly trading the W4500 (commission free), everyday, from May 2004 through May 30, 2014, BUT, based upon a variable, pre-programmed forced WINNING PERCENTAGE. As I increased that percentage (by 0.1% at a time), the frequency of having a "losing decade" decreased.

But I was looking for the "holy grail" percentage number. A number where, even when running the simulation scores of iterations--with random trades, I could not find a losing session. That number is 55.1%.

Setting the experiment to trade randomly, but with a 55.1% winning rate; I was unable to produce a losing trading session after scores of iterations.

 

[ebbnflow]

Re: Still Clueless?

Ebb,

FWIW

I ran the experiment. The experiment simulates randomly trading the W4500 (commission free), everyday, from May 2004 through May 30, 2014, BUT, based upon a variable, pre-programmed forced WINNING PERCENTAGE. As I increased that percentage (by 0.1% at a time), the frequency of having a "losing decade" decreased.

But I was looking for the "holy grail" percentage number. A number where, even when running the simulation scores of iterations--with random trades, I could not find a losing session. That number is 55.1%.

Setting the experiment to trade randomly, but with a 55.1% winning rate; I was unable to produce a losing trading session after scores of iterations.

Userque, maybe my strategy is right after all by going long only if the pattern's win percentage is over 55.1%.

 

[userque]

Re: Still Clueless?

Userque, maybe my strategy is right after all by going long only if the pattern's win percentage is over 55.1%.

Yep, I noticed that too (as far as your long trades go). The flip side of 55.1% is 44.9% (for going short). But I would rather do another experiment separating the long trades from the short trades--just like you do. The original experiment used the same percentage for both long and short trades. I suspect the numbers won't be exactly "flipped" in the next experiment as the market behaves differently going down, versus going up.

Keep in mind that the non-forced trades (the non-55.1%) were random. If those trades were not random, but were 'attempts' to be correct as well (as your trades are), then a percentage lower than 55.1 would probably be optimal.

These experiments may reveal upper and lower bounds, but optimal numbers can likely only come via backtesting your system using different parameters, or "forward" testing your trades in a spreadsheet as you go along, day by day--doing what-ifs with different parameters after each session. The latter is easier for the non-spreadsheet guy but will take a long time to get an accurate number; the former will get you their instantly, but requires much more coding.

Sounds like you're doing something like the latter.

Bottom line is: it works!

 

[ebbnflow]

Re: Still Clueless?

End of month recap: My TSP had a gain of 3.00% in April, and 4.12% this May, so the TSP system is now up +9.78% on the year. Still advancing steadily and with two new IFTs available, I will be on the lookout for another double-pattern entry (current winning streak is 7).

 

[ebbnflow]

Today's market action has pattern 8/red-grn-red written all over it. It was a perfect snapshot (Kodak moment). :nuts:

Friday: Pattern 8/red-grn-red. Win Percentage (CSI 54.9%): C 55.6%, S 52.3%, I 56.8%.
Result: long C-fund was right; short S-fund was right; long I-fund was right.

Unofficial (May 30, 2014): C +0.18%; S -0.30%; I +0.05%.

Short Patterns (S-fund): 2/grn-grn-red (51.9%); 8/red-grn-red (52.1%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%); 7/red-grn-grn (53.1%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.1% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.1%; long > 55.1%.
Accuracy: 23/34 (68%). Long: 16/22 (73%). Short: 7/12 (58%). Cash: green (3 up, 3 down); red (3 up, 3 down).

Monday: Pattern 2/grn-grn-red. Win Percentage (CSI 53.3%): C 55.1%, S 51.9%, I 53.0%.
Forecast: long C-fund; short S-fund; short I-fund.

Monday has pattern 2/grn-grn-red. Win Percentage (S-fund): 51.9% (W 123, L 114). Strategy: Short <= 53%.

Almost had a negative S-fund (+0.00%), but TNA (-1.57%) did tumble, so it was the right call.

Monday: Pattern 2/grn-grn-red. Win Percentage (CSI 53.3%): C 55.1%, S 51.9%, I 53.0%.
Result: long C-fund was right; short S-fund was wrong; long I-fund was wrong.
Unofficial (June 02, 2014): C +0.07%; S +0.00%; I +0.18%.

Short Patterns (S-fund): 2/grn-grn-red (52.1%); 8/red-grn-red (52.1%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%); 7/red-grn-grn (53.1%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.2% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.2%; long > 55.2%.
Accuracy: 23/35 (66%). Long: 16/22 (73%). Short: 7/13 (54%). Cash: green (3 up, 3 down); red (3 up, 3 down).

Tuesday: Pattern 7/red-grn-grn. Win Percentage (CSI 52.2%): C 54.5%, S 53.1%, I 48.8%.
Forecast: cash C-fund; cash S-fund; short I-fund.

Tuesday has pattern 7/red-grn-grn. Win Percentage (S-fund): 53.1% (W 111, L 98). Strategy: Cash > 53% and <= 55.2%.
Note: A loss in the S-fund for pattern 7 tomorrow, switches it from cash to short.

 

[ebbnflow]

Heads-up: With 2 new IFTs available, I am once again looking for a double-pattern entry. I have found one, but this time I plan on entering earlier than usual. We have a bullish pattern 3/grn-red-red this coming Friday (jobs report day). Its win percentage in the S-fund is 57%, so that makes it a long pattern. Next Monday has cash/neutral pattern 4/grn-red-grn. The following day on Tuesday, is the start of our bullish double pattern 1-1 which has a win percentage of 69% (1st day) and 64% (2nd day). The IFT will be done before Thursday noon, so I can be in the S-fund by Friday (June 6). The stay should be from June 6 to 11.

 

[ebbnflow]

Almost had a negative S-fund (+0.00%), but TNA (-1.57%) did tumble, so it was the right call.

Monday: Pattern 2/grn-grn-red. Win Percentage (CSI 53.3%): C 55.1%, S 51.9%, I 53.0%.
Result: long C-fund was right; short S-fund was wrong; long I-fund was wrong.
Unofficial (June 02, 2014): C +0.07%; S +0.00%; I +0.18%.

Short Patterns (S-fund): 2/grn-grn-red (52.1%); 8/red-grn-red (52.1%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%); 7/red-grn-grn (53.1%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.2% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.2%; long > 55.2%.
Accuracy: 23/35 (66%). Long: 16/22 (73%). Short: 7/13 (54%). Cash: green (3 up, 3 down); red (3 up, 3 down).

Tuesday: Pattern 7/red-grn-grn. Win Percentage (CSI 52.2%): C 54.5%, S 53.1%, I 48.8%.
Forecast: cash C-fund; cash S-fund; short I-fund.

Tuesday has pattern 7/red-grn-grn. Win Percentage (S-fund): 53.1% (W 111, L 98). Strategy: Cash > 53% and <= 55.2%.
Note: A loss in the S-fund for pattern 7 tomorrow, switches it from cash to short.

Pattern 7/red-grn-grn means trouble for the I-fund. With today's loss, its win percentage in the I-fund sinks to 48.6%, the lowest in the ebbchart patterns. It also switched the pattern from cash to short, so we now have three short (2, 7, 8), two cash (4, 6), and three long (1, 3, 5) patterns.

Tuesday: Pattern 7/red-grn-grn. Win Percentage (CSI 52.2%): C 54.5%, S 53.1%, I 48.8%.
Result: cash C-fund was safe; cash S-fund was safe; short I-fund was right.
Unofficial (June 03, 2014): C -0.04%; S -0.06%; I -0.18%.

Short Patterns (S-fund): 2/grn-grn-red (52.1%); 7/red-grn-grn (52.9%); 8/red-grn-red (52.1%).
Cash Patterns (S-fund): 4/grn-red-grn (53.6%); 6/red-red-grn (53.6%).
Long Patterns (S-fund): 1/grn-grn-grn (56.5%); 3/grn-red-red (57.0%); 5/red-red-red (60.4%).

S&P 500: 55.2% (7-yr. win percentage).
Strategy: short <= 53%; cash > 53% and <= 55.2%; long > 55.2%.
Accuracy: 23/35 (66%). Long: 16/22 (73%). Short: 7/13 (54%). Cash: green (3 up, 4 down); red (3 up, 3 down).

Wednesday: 2/grn-grn-red. Win Percentage (CSI 53.5%): C 55.3%, S 52.1%, I 53.2%.
Forecast: long C-fund; short S-fund; cash I-fund.

Wednesday has pattern 2/grn-grn-red. Win Percentage (S-fund): 52.1% (W 124, L 114). Strategy: Short < 53%.

 

[ebbnflow]

Some folks would have you think that just because I missed getting onboard the AutoTracker, my TSP returns are too incredible to believe. Here are the only three IFTs I made to the S-fund all year resulting in a gain of +9.78% at the end of May. Spent only 15 days in equities for maximum gain with limited risk. All three double-pattern entries were posted here in my thread before deadline.

A double pattern 8-8 (Jan. 14-15) showed up on the ebbchart for this Tuesday and Wednesday, so I did an IFT before noon today. The 1st day has a win percentage of 54% while the 2nd day has a win percentage of 70%. With the gains today, I hope there's more left tomorrow. It's only a one day move to the S-fund, so I'm back on the sidelines (F-fund) by Thursday. By the way, the last four excursions I did because of a bullish double pattern resulted in gains (+0.72%, +0.73%, +0.11%, +1.99%). Might need a little luck here.

Daily ebbchart:

There's a double pattern on the ebbcharts for next week (Apr. 08/Tuesday and Apr. 09/Wednesday). Usually, I use only the bullish red double patterns to enter the market for a day or two, but I'll make an exception for pattern 1/grn-grn-grn. Its normal 7-yr. winning percentage in the S-fund is good at 55%. But as a double pattern, its 1st day jumps to a great win percentage of 68%, and its 2nd day is not that bad either at 62%. That being said, I will make an IFT before noon next Monday, so I can be in the S-fund for Tuesday and Wednesday.

Double-patterns chart:

I will be making an IFT before noon today (Thursday), so I can be in the S-fund by tomorrow (Friday). The ebbcharts have a double pattern 3-3 from May 08-09. The 1st day of the DP has a WP of 50% (11/22), so the system remains in the F-fund. But the 2nd day on Friday has a high WP of 68% (15/22), so this becomes our entry point. Next Monday, we have pattern 4, which is a neutral/cash pattern. Then on Tuesday and Wednesday, double pattern 6-6 shows up, and both days have bullish win percentages -- 61% (17/28) and 57% (16/28), respectively. The TSP system will be in the S-fund from May 09-14.

 

[userque]

Some folks would have you think that just because I missed getting onboard the AutoTracker, my TSP returns are too incredible to believe. Here are the only three IFTs I made to the S-fund all year resulting in a gain of +9.78% at the end of May. Spent only 15 days in equities for maximum gain with limited risk. All three double-pattern entries were posted here in my thread before deadline.

Yeah, I just wasn't going to say anything and simply just let the minds of others work as best they can on their own, but.....the belief that either you are on the autotracker, or it didn't provably happen is limited thinking at best. There's a whole big world out there kids!

As a former website administrator myself, Tom has a well run site. Even so, this is not the only site with a tracker specific to the TSP, as nnuut has pointed out in the past. There are also reputable non-TSP based sites where one's trades of the S-Fund (etc.) underlying can be logged and viewed publicly.

Even your trades in the Challenge are questioned by those same mindsets, even when they are posted in advance for all to see, look-up, reference, and verify!

Finally, if you have a question about someone's motive, just ask. I recall having the same concerns. I went straight to the horse's mouth and guess what...asked. Simple.

You don't have to believe everyone, but before speculating negatively, why not ask first? And if you still want to speculate negatively after that, at least include the response of the person you wish to speak negatively about.?

I think the wikipedia link I posted a day or two ago explains it best.

 

[userque]

What If Ebbcharts traded the TNA instead of 3X S-Fund? See charts below:





* Please note that the lower two charts have overlapping dates.