Trading Lessons/Insights From Coin Flipping

Quote from intradaybill:

Back to the issue. The simple rule is a profitable system and your introduction of a random entry condition turns it unprofitable on the average.
More...

Actually, that's not what happens...

Quote from intradaybill:
Mike I do not like people who do not have a point in advance but try to discover what the issue is by insulting others in the meantime. You are not sure of your own conclusions and your results even disproove you. You have not made a clear statement of what you are trying to demonstrate. This is because you are not sure what you are trying to demonstrate.
More...


If you were able to step outside of your own ego, then maybe you'd realize that I have come to my own conclusions and do not want to influence any new ideas. Discussions, the way you choose to have them, are one-sided. I.E. intradaydouche dictates everything. I'd rather learn form others as they come to their own conclusions.

You see in-douche, I don't have any pre-conceievd notions of how things need to be explained or, that they have to be a certain way. I'd rather hear any and all ideas first. It's called learning by experiement and discovery. Try it sometime.
 
Quote from Mike805:

You see in-douche, I don't have any pre-conceievd notions of how things need to be explained or, that they have to be a certain way. I'd rather hear any and all ideas first. It's called learning by experiement and discovery. Try it sometime.
More...

If that were true, you would not get that defensive Mike. When people express their opinion and they are different than yours you get very aggressive. This is proof you are a dishonest individual. You are trying to discourage people who believe genuing trading systems can be developed using simple rules because you have a hidden agend of your own. I do not know what your problem is but it seems to me you are a loser trader and you are trying to justify your inability to exercise discipline and prudent risk and money management using some childish random runs. This is the most probable case with you Mike. As you see, I do not add a vulgar prefix or suffix to your name.
 
Quote from intradaybill:

If that were true, you would not get that defensive Mike. When people express their opinion and they are different than yours you get very defensive. This is proof you are a dishonest individual. You are trying to discourage people who believe genuing trading systems can be developed using simple rules because you have a hidden agend of your own. I do not know what your problem is but it seems to me you are a loser trader and you are trying to justify your inability to exercise discipline and prudent risk and money management using some childish random runs. This is the most probable case with you Mike. As you see, I do not add a vulgar prefix or suffix to your name.
More...

How in the world did you get to those conclusions? I am not trying to discourage anyone! That couldn't be further from the truth... This is why I give you a hard time bill, your conclusions are just wrong and you will never admit it... that, and you call too many people stupid.

I seriously cannot believe you think that I am trying to discourage people. The opinions being expressed by certain individals dismiss the element of randomness to a fault. Those ideas I discourage because randomness is important...

Of course geniune trading systems can be developed using simple rules... Where in the hell did you get me saying anything to the contrary?

Jeez, bill, its like you're on another planet here.

And why do you always resort to attacking my trading records? I've been doing this a long time and there are quite a few people who come on this site who know me personally and exactly what my trading results are. Get a grip... come back to earth and be rational here, I'm not going to give my personal info, but I can assure you my trading is not something you want attack.
 
Quote from shortie:

Mike is posting some good educational stuff here.

You guys should try to refrain from bickering as it pollutes the thread.
More...

Are you so stupid to think that a few tests using the random() function is some educational stuff? Get serious. The guy is an idiot. He cannot understand the difference between a deterministic and a stochastic process, he is mixing the two and he tries to make other idiots who have never taken a probability course in college to scratch their empty head.
 
Quote from Mike805:

...

What about if we reverse this condition?

Buy = Cover = TimeNum()==125400 AND rand < 0.50 AND C1 < C2;

Sell = Short = TimeNum()==125400 AND rand >= 0.51 AND C1 > C2;

...
AVG: 45427.84

Here's something funny... all but one of the runs was profitable and the avg. PnL was 45k. Something is going on here... Will the results change if I did 1000 runs? Likely not...

Can we infer that time of day and 2 days of closing prices have some value in this particular example?

Mike
More...

your evidence may suggest that that prices have a tendency to revert to mean. what's interesting is that the effect is seen even after almost one day delay since you don't check the closing price of the 1555 bar.

the first question that springs to mind is whether you will see further improvement if you use C0(1555) and C1 instead of C1 and C2. i expect to see an improvement but who knows...
 
Quote from Mike805:

Let's take a look at 25 runs with the entry conditions as posted earlier. Each value is a resulting PnL. The AVG is the average of 25 runs:
No Condition

42325
-8275
25975
111700
-287
11650
32150
-29400
10025
7100
-27450
-17237
9112
2975
23400
-34912
38725
-4775
34587
-5812
-32262
16950
-30212
105037
-2225
AVG 11154.56

Ok, so does anyone have any conditions that might affect this coin flip entry?

I'll start off with a couple:

C1 = Yesterday's daily close.
C2 = Day before yesterday's daily close.

Buy = Cover = TimeNum()==125400 AND rand < 0.50 AND C1 > C2;

Sell = Short = TimeNum()==125400 AND rand >= 0.51 AND C1 < C2;

25 Runs:
C1 > C2 & C1 < C2

-71612
-95225
-95287
-37150
-30325
-37262
-49375
-27037
-95887
-95150
-31775
-6000
-56312
-53662
-26862
-17175
-20037
-95950
-43575
-37862
-22762
-75400
-54537
-68550
-18150
AVG: -50516.76

Seems that this particular conndition virtually guarantees that we lose money when combined with a coin flip entry.

What about if we reverse this condition?

Buy = Cover = TimeNum()==125400 AND rand < 0.50 AND C1 < C2;

Sell = Short = TimeNum()==125400 AND rand >= 0.51 AND C1 > C2;

25 Runs:
C1 < C2 & C2 > C1

97087
110500
39300
22025
27175
90937
47662
29675
50837
48050
16125
91800
78362
15575
45450
67662
7687
-20150
51925
23512
43725
79475
36325
33075
1900
AVG: 45427.84

Here's something funny... all but one of the runs was profitable and the avg. PnL was 45k. Something is going on here... Will the results change if I did 1000 runs? Likely not...

Can we infer that time of day and 2 days of closing prices have some value in this particular example?

Mike
More...

Without knowing the various stats for the runs, it's hard to infer anything concrete, but on the evidence presented it seems something is going on, it would be interesting to know what happens when the random component is removed. Usually when one flips the rules, all one gets is more randomness, so I guess this could also be interpreted as a signal something non-random is happening.
 
Quote from Craig66:

Without knowing the various stats for the runs, it's hard to infer anything concrete, but on the evidence presented it seems something is going on, it would be interesting to know what happens when the random component is removed. Usually when one flips the rules, all one gets is more randomness, so I guess this could also be interpreted as a signal something non-random is happening.
More...

When the random component is removed, C1 > C2 for long and C1 < C2 for short has a win rate of about 54% and a profit factor of about 2. This shows that there is certain high degree of autocorrelation in the data. If you look at a daily chart of ES for the period he considered you will notice two large almost identical cycles, an uptrend followed by a downtrend and then another uptrend to the same levels followed by another downtrend, Thus, the deterministic signal C1 > C2 for long and C1 < C2 for short takes advantage of these long term trends to profit. When one adds a random component, as he did, the autocorrelation is no longer of an advantage, it is diffused in a way, and negative performance results.
 
Quote from intradaybill:

When the random component is removed, C1 > C2 for long and C1 < C2 for short has a win rate of about 54% and a profit factor of about 2. This shows that there is certain high degree of autocorrelation in the data. If you look at a daily chart of ES for the period he considered you will notice two large almost identical cycles, an uptrend followed by a downtrend and then another uptrend to the same levels followed by another downtrend, Thus, the deterministic signal C1 > C2 for long and C1 < C2 for short takes advantage of these long term trends to profit. When one adds a random component, as he did, the autocorrelation is no longer of an advantage, it is diffused in a way, and negative performance results.
More...

Your reading comprehension of the experiement and the results is very poor. You get an F+ for reading comp (the + is on participation points) .

There were 3 examples:

(1) No condition, i.e. time + flip
(2) time + flip + (long AND C1 >C2) + (short AND C1 < C2)
(3) time + flip + (long AND C1 < C2) + (short AND C1 > C2)

At least get the problem right before you start calling everyone else an idiot for the 100th time.

We can get into autocorrelation later. To imply that I don't know the difference between deterministic and stochastic process is very short sighted. This is an experiment bill.

To everyone else, let's keep the comments going. I will get to the good responses soon.
 
Quote from Mike805:

At least get the problem right before you start calling everyone else an idiot for the 100th time.
More...

I don't see what I missed. I said that removing time + random produces positive performance. Before you agreed to that.

Do you want to stay alone in this thread? If everyone leaves, you got no chance to learn anything.
 
You must log in or register to reply here.
Back
Top