New Forum Address: ROULETTELIFE.COM
  Update your Bookmarks

Author Topic: why would someone test a method using simulation software?  (Read 1528 times)

0 Members and 1 Guest are viewing this topic.

Sputnik

why would someone test a method using simulation software?
« on: June 04, 2018, 10:05:23 AM »


Don't understand, why would someone test a method using simulation software.
Such information after such test is useless, my opinion.

I know by hand testing if I will make money or not.

There is no point to simulate a method for several hundred thousand trails.
All system fail such test and is not how you win.
 
The following users thanked this post: mr j

Mike

Re: why would someone test a method using simulation software?
« Reply #1 on: June 04, 2018, 11:43:34 AM »
I know by hand testing if I will make money or not.
Sorry, but that's just dumb.

You might as well ask why someone would test by hand when they can write a simulation. A simulation is faster and it's just as easy to test a million spins as a 100 for the same effort.

Why is testing manually better than writing a simulation and why is testing over a million spins a waste of time? If all systems fail after a million spins doesn't it show that systems don't work?  ::)
You're just deluding yourself if you conclude you have a winning system if it makes a profit over a few hundred spins. The fact that no system can stay ahead after a million spins just proves that you have no edge but are relying on variance. Either you have an edge which can be calculated or logically inferred, or you should test a system at least to the point where the expected loss overtakes the variance. If the system continues to make a profit past this point, then you have winner ( a real edge), otherwise it's just luck.
 

Trilobite

Re: why would someone test a method using simulation software?
« Reply #2 on: June 04, 2018, 12:24:28 PM »

When this forum was young there were problems with pm's. I have never been into sending Pm's, but as a rule always try to respond to pm's I receive.

Spindizzy tried to communicate with me via pm about coding, but because of forum glitches, never received my responses.

So my guess is that he thought I was rude not to respond to his important messages, even though I did. And God knows who else didn't respond, so he left the forum thinking we were all uncouth.

 

mr j

Re: why would someone test a method using simulation software?
« Reply #3 on: June 04, 2018, 01:00:06 PM »

Don't understand, why would someone test a method using simulation software.
Such information after such test is useless, my opinion.

I know by hand testing if I will make money or not.

 

I agree.

Ken
 

GIAJJENNO

Re: why would someone test a method using simulation software?
« Reply #4 on: June 04, 2018, 01:01:37 PM »
He just does not delete his Pms, I think.
 

Mike

Re: why would someone test a method using simulation software?
« Reply #5 on: June 04, 2018, 01:30:02 PM »
@ Sputnik and Ken,
HOW do you know whether a system will make money or not, just by hand testing?
So say you've done a few tests manually and you've decided that the system will make money. WHY is it then a waste of time to test over a million spins?
I don't expect any straight answers...

 

mr j

Re: why would someone test a method using simulation software?
« Reply #6 on: June 04, 2018, 01:37:55 PM »
I cant speak for him, so I wont. I usually do around 4k hand tested spins from C. Casino. Spin by spin by spin.
I dont need NOR want 839 million spins tested. That goes for MOST members here but you wont hear that from the other cowards hiding. At least I have the nuts to talk about it. I wont play for 839 million spins so why test it as such? I dont test on a 0 wheel because we dont have a single 0 here at the casino, why test on it? etc.

Ken
 

mr j

Re: why would someone test a method using simulation software?
« Reply #7 on: June 04, 2018, 01:51:15 PM »
I wont play for 839 million spins so why test it as such? I dont test on a 0 wheel because we dont have a single 0 here at the casino, why test on it? etc.

 

In other words, if its something I wont be doing, why should I test it UNDER those conditions?

Here is another smaller example. Its sometimes difficult getting many chips down ON TIME if you play many numbers. Sooooo, I dont test/play UNDER conditions that will not apply to me.
 

Mike

Re: why would someone test a method using simulation software?
« Reply #8 on: June 04, 2018, 03:05:00 PM »
Ok Ken, so you test  your new system over 4k spins. Presumably if it does great it's declared fit for active service, otherwise it's binned.

But why 4k spins? it's an arbitrary number. A system betting only a few numbers (say 2 or 3) could EASILY do well over twice that many spins but tank and never recover over the next 8k spins. Variance depends on how many numbers are bet. High variance systems (few numbers) will on average last a LOT longer and give you more opportunities to quit with a profit than low variance systems (many numbers), which is why I agree with you that betting 2-4 numbers is the way to go.

Quote
In other words, if its something I wont be doing, why should I test it UNDER those conditions?

It's not really a "condition" in the same sense as your other example of betting many numbers and having to get many chips down in time.

Think about this : you've been playing roulette for many years now and your methodology is VERY different from what it was in the early days. How many spins (testing included) have you played over that time? I'm betting it must be several hundred thousand at least. My point is that you have LEARNED something which you probably wouldn't have if you had only played/tested over 4k spins all those years.
And you could ask, WHY NOT simulate over a million spins? If someone offered to code your system don't tell me you wouldn't be at least curious how it performed in the long run.
 
The following users thanked this post: mr j

mr j

Re: why would someone test a method using simulation software?
« Reply #9 on: June 04, 2018, 03:18:02 PM »
You are one of FEW people on this board that makes sense. Huge respect for you and I dont say that lightly.

But why 4K spins? Because its better than testing it at 200 spins (too few). You are dead correct regarding, I have LEARNED a thing or two over all those spins, tested/played......which lead me to playing fewer numbers, no progressions etc.

So this is a discussion of why 4K spins? Its a nice round number, it works fine. There is another reason WHY I test. I have done threads on this subject before. Testing is NOT only to see, if a method holds up. It also gets you practice for all the "what if" situations that *WILL* come up at the B&M casino.
If you're gonna screw something up while testing (C. Casino), I want to do it with FAKE money.
« Last Edit: June 04, 2018, 03:26:28 PM by mr j »
 

mr j

Re: why would someone test a method using simulation software?
« Reply #10 on: June 04, 2018, 03:19:55 PM »
"Variance depends on how many numbers are bet. High variance systems (few numbers) will on average last a LOT longer and give you more opportunities to quit with a profit than low variance systems (many numbers)" >>>>  :)
 

mr j

Re: why would someone test a method using simulation software?
« Reply #11 on: June 04, 2018, 03:34:42 PM »
"If someone offered to code your system don't tell me you wouldn't be at least curious how it performed in the long run" >> Three points. This has happened to myself and a few others in the past. There are CERTAIN RULES which just cannot be coded. Nothing to do with gut feelings (nothing like that) but there is a situation where I need to choose....it cannot be coded per say.

The second point, I cannot trust the results would be FAIR. Good or bad results, I cant trust the person doing it. Also, lets say the coder loves the method. He now has a FREE kick a** method from me. That aint gonna happen.
 

mr j

Re: why would someone test a method using simulation software?
« Reply #12 on: June 04, 2018, 03:42:50 PM »
I remember we did a great thread at GG a long time back. We averaged everything.....how many years playing at a casino(s), how many days per month, how many spins per week etc. We had almost an exact number but keep in mind, it WAS an estimate.

So, we decided on xxx number of spins over a LIFETIME (I dont remember the number). From there, why not use that number for testing?
Another point, all testing was done in gaps Meaning, 280 spins were used, an AVERAGE day of playing. Then, a different set of numbers were used, on and on etc. Everything was simulated AS IF it was normal betting. No 600 million straight RNG spins. I see no point to that.
 
The following users thanked this post: Sputnik

Sputnik

Re: why would someone test a method using simulation software?
« Reply #13 on: June 04, 2018, 04:16:09 PM »


It has been many great topics at GG …

Cheers
 
The following users thanked this post: mr j

mr j

Re: why would someone test a method using simulation software?
« Reply #14 on: June 04, 2018, 04:20:57 PM »
I could not agree more.