50k spins downloaded from rouletteresearch.com:


[attachimg=1]

I was just curious to see what happens in the really long term, so I also downloaded the 1M spin file:


[attachimg=1]

Can we verify these Macao spins?

Because clearly those results appear to be abnormal. Not just the constant upswing, but the actual size of it. None of the other graphs show a return anything like that. There is a similar trend in the rouletteresearch results, but it climbs back over something like 18k spins. Not ideal is it? Just appears random. I am struggling to see the 5% proposed advantage here!

Well Bayes, what is your final conclusion on this?

I mean, does this bet differs than any other with playing 3 numbers?

Apparently it has no mathematical edge, but can we say at least variance is lower?

Cheers

Quote from: XXVV on February 19, 2014, 06:46:30 PM
I suspect Bayes is in no position yet to make any judgment as the source of data is essential to be verified and from what I am observing depending on source wildly conflicting responses.

Yeah it seemed to me too that results are somehow conflicting from few different testers and your personal results.

Bayes.....a verified set of spins

I have noticed something


In RX....if you download spins from a casino, it does 1 table at a time.


So I set the time to several months and one continuous file. (see attachment)


Also, as long as I cut off the download before todays date, I have a massive spin file from 1 casino and 1 table.


(see attachment)


Attached is a notepad file (weus20130205_20140212_2.txt) with 30K spins from table 2 Spielbank Wiesbaden Casino May2013-Jan2014


Pretty verified if you ask me


(if I set to 2010 or so, I would get 1M spins from table 2)

XXVV....as a man of my word...I haven't even seen the contents of this file.


I hope Bayes can run it.


It's a file of integrity......as far as I trust RX to do the work for me (Yes...I am trying to trust something on face value....not like me at all)

XXVV,

There are many threads opened about this now, are we referring to the Weisbaden simulation Bayes posted here:

http://betselection.cc/xxvv-studio/wf3-testing-and-variations-discussed/msg27821/#msg27821

Yes it shows a nice upswing at least from negative territory from about -100 at spin 2000 to +600 at spin 3800. (I am looking at WF4 here, but the trend with WF3 is similiar anyway). And yes the end results are fairly positive. But lets interpret the behaviour a bit instead of just looking at the lines.

I am not sure what the projected edge is supposed to be (I am figuring 5% from a previous comment of yours). But the graph is not indicative of that. The bet is still negative after 2000 spins, which is clearly a concern considering there is supposed to be a mathematical advantage.

Once the upswing takes place and the BR rises to around 600, there is no more positive gain made after that in the remaining 6200 spins. Bear in mind I am estimating the interval of spins here! Nor am I trying to shoot this down. It seems reasonable to voice opinion for anyone who is thinking of investing real money in this.

The simulation does end positive, but no gain in more than 6000 spins suggests the bet was simply treading water and therefore was probably enjoying a period of positive random fluctuation.

As for the Macao spins file, the results look suspect to me. I understand you have tested those spins manually, or some of them. In my opinion manual testing is pointless because it can lead to false conclusions. The reason is, with manual testing, you always know the outcomes in advance. It then becomes too easy to make mistakes, make ad-hoc adjustments to the bet and ultimately begin curve fitting the results. This can all happen subconsciously. It appears that WF3 was practically tailor made for that spins file. And now against outcomes from other sources, it's struggling. I don't accuse anyone of doctoring results or moving the goal posts. It just happens. I have fallen for it myself in the past.

Undoubtedly TwoCatSam has the best method of testing systems and progressions by having them "sheeted" and run through ExcelBot. But credit to Bayes for the simulation.

Hi Number 6


I can assure you that the WF3 & WF4 testing on the Macao spins was done through an excel sheet with 100% accuracy and therefore no mistakes of the human kind


It took me many hours of manually scanning and double checking the numbers before converting them into a usable spin file from the authenticated booklet that came with the Roulette for the Millions book


Hope that answers your query


Gordon 

I do not understand why so much hue and cry over this idea. It obviously can not have an edge, in long run. No method can filter out better numbers to bet and can win flat bet. If it is even remotely possible without any MM requirements,it would kill casinos.
                                      Old ideas wrapped in new names do not make a winner. Without a classy MM, all betselection will lose.

Quote from: Albalaha on February 20, 2014, 03:57:28 AM
I do not understand why so much hue and cry over this idea. It obviously can not have an edge, in long run. No method can filter out better numbers to bet and can win flat bet. If it is even remotely possible without any MM requirements,it would kill casinos.
                                      Old ideas wrapped in new names do not make a winner. Without a classy MM, all betselection will lose.

So, just to clarify, you're basically suggesting XXVV is a liar, are you?

In fact, Albalaha, why don't you come right out with it and actually call  him a liar?

XXVV,

It doesn't mean to be offensive, it's a mere observation that in manual testing, results can easily be corrupted. Read around anywhere, time and again you'll see people complain of rigged RNGs. They hand test for a few hundred spins and think they've got the grail. They lose first time with real money, and think they must have been cheated. You'll see people like John Legend who also hand tested his systems. He thought he would change the world with a martingale. When he engaged in some meaningful test with Superman, he lost his BR almost instantly and has never been heard from since.

Obviously I believe your tests to be more comprehensive and sensible, but regardless corruption is still possible. The Macao spins show a clear reversal compared to outcomes sourced from other places. The question I pose is pretty simple: can those results be trusted? Not, as in, have they been fabricated? Or has the bet been reverse engineered to beat those spins?

It just means, in the grand scheme of things, can we rely on them when gauging WF3 or analysing it? Because the nature of the trend is very different. It could be a real edge in play, or could be a fluke. Those results are at odds with others. So, can Macao be discounted for some reason?

For now I would run Turner's Wiesbaden download through the simulation. If the results are positive it would warrant another round of testing.

Either way, you're correct I think the testing is become too divided and slapdash. I do think "official" tests should be done with spins from a verified source.

Quote from: XXVV on February 20, 2014, 06:44:02 AM

The arguments and presentation of rogue sourced 50K spins or worse 1million spins is so out of touch with what is real, valid or worthwhile it astonishes me, and I would have thought the individuals concerned would have had better judgement. Real play is in the context of 1-say 200 spins ( 3 hours +). That is the foundation of our success as players reading short cycles.

Actually I don't care what you find in your testing because in some cases as Mr #6 so smartly observes 'subconscious conditioning ' can affect preparation of results and distort findings. Well, that principle applies to testing as well as writing or interpretation or review. We cannot escape it. If you are a cynic you will perceive the world with that mindset.

We have seen some of this.

XXVV,

Oh dear oh dear... Where to start?

In the first place, how do you come to the conclusion that the spins from RR are "rogue sourced"? and what does that even mean? By "rogue" you're implying that they are biased or corrupted in some way I assume, but where is your evidence for this? simply that the simulator generated a negative result?
Why would the owner, who, judging by the quality of his site and products, is a serious researcher, put these spins in the public domain if he knew they were corrupted? Furthermore, if they are biased they can only be so with respect to some number/group of numbers or system. But without knowledge of what system the spins are to be tested against (which is impossible, given that anyone can download them), it would be pointless. The only "universally biased" set of numbers (in the sense that no system will have an edge with respect to them) are those which are generated by an unbiased wheel, which is why the casinos make such efforts to keep the wheels as random as possible.

Not only that, but I'm sure the owner would have put the spins through one of his own testers to check for corruption before uploading them. For those without such proprietary software, there is the Dieharder suite of programs which are in the public domain. It's worth mentioning that the 1 million spin file in question has been available in this forum and rouletteforum.cc for several years, and no-one has complained about them.

And as for your statement that "real play is in the context of 200 spins", well I can't believe you're dragging up such nonsense. Quite apart from the absurdity of "hit & run", may I remind you that you yourself stressed more than once that WF3 should be tested against at least 20,000 spins, in order to eliminate any short term fluctuations.

You can't have it both ways.

I'm not sure what the take-away message is from your second paragraph above. You seem to be saying that objectivity is not possible because of the dangers of 'subconscious conditioning' and 'interpretation'. Again, this is hogwash. If that really were the case, it would apply as much to your positive Macao results as anything else, so those results would also be worthless on your own terms. If what you suggest is true, then no scientific progress would be possible and we would flailing hopelessly around in a sea of skepticism. Yes, there are real dangers such as #6 highlights, but it's for that reason that statistical and scientific methodology has been developed over many years; anyone with a background in a STEM discipline would understand this.

I find your choice of words curious: "Actually I don't care what you find in your testing". It comes across as rather arrogant and with a touch of the petulant childishness you've sometimes displayed at other times when things don't appear to be going the way you'd like them to. Does it not occur to you that others DO care what the testing reveals? I for one care, and I hope others do too.

@ Turner,


Many thanks for gathering and uploading the Weisbaden spins. I'll post the results shortly. 

lets say it in short WF system is not better than any other method  no edge gained or whatever ...Bayes great dedication and thanks for your time coding it.....