Do you make these mistakes?

[Steve]

Here's an outline I published about how my knowledge developed. It's part of an article i recently published. There's a lot more to the stry. but even from this part, does it remind you of yourself?...

The Typical Stages of Learning

Almost every professional player goes through the same stages of learning. I'm no exception, so I'll explain how my knowledge progressed.

The very first time I played roulette, I observed the winning number display and tried to guess which number or color was "due". It looked like there were all kinds of distinct patterns. For example, streaks of reds, streaks and blacks, and sometimes an equal numbers of reds and blacks. It didn't seem random.

What I should have done is started by researching what is known to not work, and why. It would have saved me time and money. But I just wanted to learn for myself without anyone's input. I attempted to justify my wilfull ignorance in various ways.

I eventually learned the "patterns" weren't as simple as I thought. I believed the patterns I was seeing were real, but I just didn't understand them yet. At that stage, I didn't understand I was just seeing normal random spins, and I was making up the patterns in my head. They didn't really exist.

For some time, I constantly updated "what I thought I knew", based on how my new system performed. If I won $500 or so, I would be convinced my system "worked". And later when I lost, I would feel disillusioned, and just thought I had to tweak some system rules to "get it right". Looking back now, I know I was clueless, and my approach was all wrong . . . but I didn't think that then. And I wouldn't have believed anyone who told me I was clueless. After all, I was "mostly winning", so why would I believe otherwise?

Eventually I settled on what I felt was my "master system" that seemed to almost always win. It was basically betting on dozens with a progression. And for around a year, I was clearly profiting. I probably won over about 8,000 spins in 12 months. So why wouldn't I think my system "worked"?

But eventually, I started to lose. I couldn't imagine why, because I was winning before. I didn't know why I was losing. I tried tweaking the system rules more, but still kept losing. I even thought the casino might be cheating.

I then started more extensively testing my system on roulette simulators on my PC. I have no idea how many different systems (and variations of the same approaches) I tested. But I would basically think of an idea for a system, then test it as a "complete" system. I wouldn't test just the "principle" that is supposed to make it effective. I tested the whole system in manual simulations. This was extremely time-consuming. A test of 1000 spins or so took hours. Sometimes I won, sometimes I lost, and I didn't understand why. All I did was just "tweak the system rules" more, and hope that I finally got it right.

And when the test results were good, I'd bet for real. I would win for a while, then lose. It was incredibly frustrating. But for some time, it never occured to me that my approaches were all wrong. I just thought I was missing some small rule to make my system perfect. I always believed I was very close to discovering the secret.

To make matters worse, it wasn't as if I was uneducated. Actually I was quite well educated in mathematics, statistics, and physics. I was just young, impatient and wilfully ignorant. It wasn't that I was stupid. But wilfull ignorance has the same effect as being stupid.

Becoming Disillusioned

I began researching systems that do and don't work (at least historically), and focused on the explanations. I came to understand my entire outlook had been wrong. I was looking in the wrong direction. I began to learn various methods based on physics. I saw that they worked because I could plainly see the physics involved, but I wanted an "easier" way. Considering I've always been interested in physics, and even studied applied physics in college, you'd think physics would be the first thing I'd try. Perhaps I just wasn't thinking. The biggest thing I overlooked was my bet selection didn't change the odds of winning. Specifically with my dozens system, the patterns I thought were happening actually never happened. It was all just normal random behavior. My odds of winning on a dozen, for the next spin, were always 12/37.

For some time, I dismissed physics because I didn't want to collect data and look at things like wheel behavior. It seemed too tiresome. Tracking things like streaks of winning numbers and repeating numbeers just seemed easier and more appealing.

Properly Testing

I came to understand manual testing systems would take far too long. So I designed software to automate part of my testing. It took a few weeks to have it coded by a programmer. Eventually it was capable of doing years of manual testing at the click of a button.

To use the software, I would first define a pattern I thought existed. For example, I could test the theory that after a streak of perhaps 9 REDS, BLACK would be "due". But I found the chances of red or black spinning were exactly the same. I could test this for billions of spins, and the results were clear. There was no limit to the complexity of patterns I could search for, so I tested everything I could think of. And the result was NONE of the patterns I thought existed were real. The odds of me winning hadn't changed at all.

This all meant the results I got with my systems were no better than random. But your mind plays tricks on you in gambling. I had convinced myself that the winning sessions were proof my system worked. And the losing sessions just meant I needed to tweak some small system rule I was missing. But now with my software, I knew I had been kidding myself.

I further developed my software to apply a kind of artificial intelligence. Specifically it would randomly define a pattern, then search for it in random spins to see if the pattern might exist in reality. For example, it would randomly choose to check how often 32 would win if 32 appeared in the last 10 or so spins. And it would check over hundreds of thousands of spins (from a random number generator).

Basically I found some patterns the software found appeared to really exist. Buit then I tested over more spins, and the pattern didn't exist anymore.

I later came to understand even over hundreds of thousands of spins, some "events" are probably going to happen more than others. But that's just random variation, known as "variance". It's similar to how you can get more reds than blacks in 10 spins. But it's meaningless and doesn't apply to the next 10 spins.

Sometimes I would let my software run for many days, constantly searching for patterns in spins. Ultimately what I found was:

No matter what pattern I thought of, the odds of any number spinning next or anytime soon had not changed.

In the rare cares where patterns did appear legitimate, they were the result of poor quality "random number generators" (you dont find them in modern casinos).

Even in high quality random spins (RNG), there do appear to be clear "fractals". But they appear in a way that are retrospective, and cannot be used for future predictions.

--

[Steve]

Here's more (one of the major turning points for me):


The Switch To Real Spins

Up to this point, almost all my testing had been done with RNG spins (no real wheel, just random number generator spins). I started using my software to test real spins.

The first challenge was I couldn't get enough data for proper testing. I learned from previous tests that even hundreds of thousands of spins can give you an inaccurate understanding of what patterns are real, and which are just random variance. I couldn't just test 1000 spins or so, because the software was highlighting countless "potential patterns". And I knew it was more likely a result of having not enough data. Basically anything can happen in the short-term. You can only see the bigger picture by checking more spins.

At the time, it wasn't realistically possible to get millions of real spins. So I had to find a way of confirming the patterns the software found were legitimate, and not random variance. And it needed to be done in far fewer spins.

The solution I found was to modify the software so it could check for patterns, but only with specific variables.

Initially I collected data directly from my local casino's wheels. It was extremely time-consuming, but my local casino had wheels that were spun once every 45 seconds. I was able to collect about 300-500 spins per day for testing. I was very careful to ensure nothing about the wheel changed on different days. Usually the same ball type and size were used. Sometimes the size varied, but I just noted the size used for each group of spins. There were many other variables I tracked - basically everything imaginable, because I didn't know what was relevant.

When I entered the data into my software, the real patterns began to emerge. I further modified the software to "cross reference" the variables and patterns. Ill explain basically how this works . . . If for example a pattern was found on a wheel, the definition of the pattern is saved to memory. Now if the "variables" change, we expect the patterns to change. Or do they? . . . I eventually found some patterns do change, but others stay the same - but are unrecognizable if you don't know what to look for.

This led me to modify the software to:

* Identify statistical anomalies,

* Link the anomalies to the physical variables of each spin,

* Model the dynamic and ever-changing relationship between the variables, other variables, and the winning number.

. . . and to do it all within much fewer spins, so that real application can be practical.

This ultimately resulted in a system that really worked. All that was left to make it practical was for the software to produce convenient betting charts I could take into the casino, so I knew when and where to bet.


Summarizing The Software

In simple terms, my software looks at CAUSE AND EFFECT.

It is not exactly a "roulette system". It does statistical analysis to find anomalies in spins, then uses additional data to "cross reference" the data, to verify the patterns are likely legitimate.

Importantly, instead of roulette being completely unpredictable, it is in fact quite predictable once you know what to look for. And there are numerous legitimately effective ways to beat roulette. I discovered a few ways. I find some are great for some casinos, but not for others - you may find other even easier techniques may be better. However, there is such a system that is clearly best for MOST casinos. And this is the main system and software I explain on my website. But just in case it is not the best system for your casino, I also include almost every other legitimately effective technique. This ensures that no matter where you play roulette, the chances are you'll be able to beat your local casino.

If you want recreate my software, you can do it. You dont need my software. You just need to understand and apply what I've said. You are looking for statistical anomalies, and cross reference them with other data for verification.

------

I'll just reinforce the fact that there's no one system that works everywhere. All AP needs to be adaptable. My system software can be useless on some wheels, but perfect for others. In some cases, even basic VB can be better. It just depends on the wheel and conditions.

And no, I'm not referring to roulette computers. You dont need a computer to beat roulette. Quicker, easier, more profitable, more fun. It is just my preferred method. If you had a Toyota and a Lamborghini, which would you drive?

I know many of you are looking for the magic bullet that wins everywhere. Good luck, maybe you'll find it. But FFS try something NEW. Not the old bullshit. Dont you think it would be helpful to understand what others have tried and tested to exhaustion, and know to fail?

Many players fail when it comes to testing. They think a silly 1000 spin test can tell you anything. With AP and real variables, it can and does. But with RNG and nonsense like repeaters, it cant. You need statistically significant data.

If you need proof, use Roulette Xtreme and test a random system for 10,000 spins. Sometimes you win, sometimes you lose. Or better yet, test each system one by one, for 10,000 spins each. You'll find some actually win over 10,000 spins.

Now take that system and go play it for real. You'll probably lose.

Then go back to your PC and test that same system for another 10,000 spins. You'll probably lose.

Think about it.

Did the system magically turn bad. Or was it always a loser, and your initial spin was just short-term variance?

You guys really need to understand this point. Some of you are seriously thickheaded. Don't get me wrong. I'm not saying I'm better than you. Because Ive been where you are. I was thickheaded for years.

If you want to tinker with system for fun, whoopie, nobody cares. Have fun. But if you intend to take roulette seriously, pay attention to what I've said.

[ati]

Good post Steve, I'm sure most of us made the same mistakes.
Who cares if people say Martingale doesn't work? When I began to play, I ran simulations and the chart almost always went up to the sky in a straight line. So I thought starting with 1 cent units I can survive 10-15 losses in a row.  Well, you can guess what happened after. 

I also tried betting just 1 number with progression. Sure I won 20-30 days in a row, then once my chosen number didn't come up for over 200 spins. I was betting my monthly salary on a single number at that point. And it happened more than once. I lost a lot and in my anger I came to the forum to rage 

After that I started to create more "sophisticated" systems, based on repeat statistics, patterns and what not. Obviously didn't work.
Then I discovered non random and cycles.... I know our opinions do not match on this, but from that point I stopped playing for real money, so if nothing else, it saved me a lot of money and disappointment. There is more but I stop here to avoid unnecessary arguments. 

[nottophammer]

How's your pre-cog going

[Steve]

Ati, agreed on first part of what you said. While learning its a rollercoaster when you think you know, then finding youve got no clue. Until you finally at least understand some critical basics like statistical significance, variance, odds, payout, proper testing etc. At least the basics

Regarding the last part you said, if you have information that backs your belief, if be inclined to believe it, unless I had even more and better information to the contrary. Truth isn't an opinion, so one of us would be wrong. A logical way to determine that would be ample testing, so variance isn't an issue.

Notto, of you're asking me about precog, I've made my beliefs clear. Basically I find there's ample evidence to support there something in it, but I don't know for sure. More work is needed on it.

[Proofreaders2000]

Gambling is a team sport imo, so before
you go near a table, slot or bookie-get partner(s).

Does he/she have to be knowledgeable?

It helps.  As long as you and your partner(s) are willing to work
hard to research all aspects of the game(s) you wish to play, ok 

A tentative template:

While all spins, card game (a blackjack hand) are independent trials. 
I've found it fruitful to look for connective elements bt each game.

Example: 20, 8 (newest spin-value)

Both are Black, Even, Middle column, nine
neighbors from one another (Euro wheel)

Find a way to quickly group numbers (not a system per se, but helpful imo)

Just watch that group of numbers (could
be a dozen or a six numbers to scrutinize)

Record what happens (you never know what you see)

*I hope this inspires ideas (or at least some discussion). 

[Steve]

Proofreaders, it's as if you completely ignored everything in my post, then went ahead and made the same mistakes I was talking about.

Blacks, evens, zero, dozens, repeaters, hot numbers, cold numbers blah blah blah, it's all a deliberate distraction to keep you looking the wrong way.

[Proofreaders2000]

I am simply adding my input to the discussion.

Since each game is an independent trial
why not make the most educated guess possible?

[Steve]

Educated is good. We already know looking at reds, blacks etc doesn't work.

[Roulettebeater]

Steve
What you wrote down is good but not always possible, I ll tell you why.

I personally have a system and I think it hs an edge but I can’t perform “back office” tests !

What I mean with back office tests ?
Simplly are tests / evaluations that can be done anytime.

My system unfortunately requires three inputs :  launch point, exit point (where the ball leaves track) and final point .

Having a saved sets of spins doesn’t help me at all, as I can’t test ! The only way to test is to log in a live casino and perform live tests, but this is a time killer process !

[Steve]

That's like most AP

[Roulettebeater]

That's like most AP

It’s AP

[Roulettebeater]

That's like most AP

It’s AP

[Steve]

That's why I said it. When you stop looking at fairytales (repeaters etc), and start looking at reality (things that do affect spins), it's AP. Whether or not the approach is optimal is another matter, but it's definitely the right area.

[Roulettebeater]

That's why I said it. When you stop looking at fairytales (repeaters etc), and start looking at reality (things that do affect spins), it's AP. Whether or not the approach is optimal is another matter, but it's definitely the right area.

But testing is a major issue !
I am used to perform chunk tests !

Any tip ? Or we should simply skip extended tests and attack the game ASAP?