Author Topic: RTM  (Read 17128 times)

0 Members and 1 Guest are viewing this topic.

Sputnik

  • Mature Member
  • ***
  • Posts: 494
  • Thanked: 390 times
RTM
« on: March 22, 2015, 09:28:48 AM »

I never forget this quote from WIKI:

"Regression toward the mean simply says that, following an extreme random event, the next random event is likely to be less extreme. In no sense does the future event "compensate for" or "even out" the previous event, though this is assumed in the gambler's fallacy (and variant law of averages). Similarly, the law of large numbers states that in the long term, the average will tend towards the expected value, but makes no statement about individual trials. For example, following a run of 10 heads on a flip of a fair coin (a rare, extreme event), regression to the mean states that the next run of heads will likely be less than 10, while the law of large numbers states that in the long term, this event will likely average out, and the average fraction of heads will tend to 1/2. By contrast, the gambler's fallacy incorrectly assumes that the coin is now "due" for a run of tails, to balance out."

So i run 1 million trails to see the worst and extreme for even money bets:

Low: 18 in a row (once)
High: 18 in a row (once)
Red: 19 in a row (once)
Black: 20 in a row (twice)
Odd: 18 in a row (twice)
Even: 18 in a row (once)

So two times in 1 millon trails, so was there no hit or regression towards the mean.
I just test this as i would like to confirm that 10 events are extreme and that our expectation for the next 10 trails will include at least one winning bet.



 

Sputnik

  • Mature Member
  • ***
  • Posts: 494
  • Thanked: 390 times
Re: RTM
« Reply #1 on: March 22, 2015, 10:26:27 AM »
19 contra 1
is 4+ STD
That explain why its so rare, when 5.09 STD appear once during 1 million trails.
50 loses and 7 wins

Now is the question if we can fool the random flow.
Fource it to show us the nature behind RTM.

All cominations or patterns has the same probability.
This means that we can pick any 10 random results and compare them whith the next 10 random results.
Then we should see regression toward the mean with several wins as lower STD is more common.

Code: [Select]
1
2
2
2
2
1
1
1
1
2

2 O W
2 S L
2 S L
1 O W
2 S L
2 O W
1 S L
2 O W
1 S L
2 S L

1 O W
2 S L
1 O W
1 S L
1 O W
1 O W
1 S L
1 O W
2 O W
2 S L

2 O W
1 O W
2 O W
1 S L
1 S L
2 O W
1 S L
1 S L
1 O W
1 O W

2 S L
2 O W
1 O W
2 O W
2 O W
2 S L
1 S L
2 O W
1 S L
2 O W

Now if our benchmark is 4.08 STD or 22 contra 2
Then our expectation would allways be at least 3 events

This mean i can compare 11 with 11
But i don't have to bet every trail as i know there will be at least three shows.

L L L W L L L W L L L

That is 11 contra 2 and my expectation is 3
As you can see i only need to attack three times with three attempts as there will be at least one more win among does trails

But in reality i might bet three times L L L and there is no win and loses chop L L L L L
Then i have to bet less to cover my probability window with my expectation of 3 events

This describe the situation with "Diversification"

In finance, diversification means reducing non-systematic risk by investing in a variety of assets. If the asset values do not move up and down in perfect synchrony, a diversified portfolio will have less risk than the weighted average risk of its constituent assets, and often less risk than the least risky of its constituent.

This also put some light on the worst as 50 loses with 7 wins means you had several sequenses with 10 following by 9 or less.
 

kav

  • www.Roulette30.com
  • Administrator
  • Hero Member
  • *****
  • Posts: 1830
  • Thanked: 801 times
  • Gender: Male
Re: RTM
« Reply #2 on: March 22, 2015, 10:42:39 PM »
Sputnik this is a great post that could fire a great discussion. Thanks.

Now I must admit that I lost you in the middle of you second post.
Let's try to make this simple and go one step at a time, like Belgian did in his thread.

So let me state what  do understand that you are saying and correct me and complete me.
  • Extreme events (deviations) can not go on forever.
  • Extreme events do not repeat themselves often And if they repeat themselves they are milder.
  • In one million spins test these were the Even Chances extreme events:
    Low: 18 in a row (once)
    High: 18 in a row (once)
    Red: 19 in a row (once)
    Black: 20 in a row (twice)
    Odd: 18 in a row (twice)
    Even: 18 in a row (once)
  • In 1 million spins the greatest length high deviation observed was 7 wins in 57 sins (5,09 STD) This is a great observation IMO. It may be actionable!
  • It is safe to assume with high degree of certainty that in 22 spins we will observe at least 2 wins.
Now that's all I got from your great posts. Please explain further. Comments from other members are also welcome. Let's try to get something out of this.
 

Bayes

  • Veteran Member
  • ****
  • Posts: 688
  • Thanked: 550 times
  • roulettician.com
Re: RTM
« Reply #3 on: March 23, 2015, 05:35:10 PM »
In my experience, the best that RTM can offer is to make things a little less volatile, but it offers no mathematical advantage in the sense that a bunch of wins is more likely following a drought, so flat betting is not an option.

The independence of trials applies to sequences of outcomes as much as to single outcomes, so if in the last 100 spins there have been only 30 reds or as many as 70, it is still unlikely that either of these two scenarios will repeat back-to-back, not because of what has just happened, but because they are unlikely whatever has happened.

So you might be wondering what the value is in only starting your attack after a drought if such extreme events are rare in any case? The point is that it does reduce the variance (the gaps length between losses) if you time your entry points accordingly, and lessens the strain on the bank (and your nerves).

I've tried both betting with the trend and against it, the latter gives measurably better results, at least where the even-chances are concerned.
« Last Edit: March 23, 2015, 05:37:22 PM by Slacker »
 

Real

  • Fighting the war on absurdity one foolish idea at a time.
  • Hero Member
  • ******
  • Posts: 1378
  • Thanked: 201 times
  • Gender: Male
Re: RTM
« Reply #4 on: March 26, 2015, 05:23:27 PM »
If in the last 100 spins there have only been 30 reds, then the random walk dictates that your expectation moving forward is expectation, nothing else. 
 

Trilobite

  • Mature Member
  • ***
  • Posts: 334
  • Thanked: 115 times
Re: RTM
« Reply #5 on: March 26, 2015, 11:02:34 PM »
Reducing volatility is a very important factor when trying to beat roulette, so in that respect focusing on RTM in some fashion should help accomplish that goal.

I find RTM reduces volatility just as well if not better when applied to the wL registry.
For example, if you are playing the EC’s then you should be winning close to 50% of your bets over the long run, but in shorter cycles of play you might be winning well above the average or well below the average. When the game trends to extremes of winning or losing you can with better than average certainty predict a breakdown of the trend to some degree, or a RTM.

When everything’s going right it’s easy to think you will keep winning forever, but you won’t. The losers will start to come at a rate that will somewhat correct your overall long term averages.
 
When everything’s going wrong you sometimes think you will keep losing forever, but you won’t. The winners will start to come at a rate that will somewhat correct your overall long term averages.
 
« Last Edit: March 26, 2015, 11:04:10 PM by Trilobite »
 

Real

  • Fighting the war on absurdity one foolish idea at a time.
  • Hero Member
  • ******
  • Posts: 1378
  • Thanked: 201 times
  • Gender: Male
Re: RTM
« Reply #6 on: March 27, 2015, 12:05:56 AM »
Quote
Reducing volatility is a very important factor when trying to beat roulette, so in that respect focusing on RTM in some fashion should help accomplish that goal.

Nope. It wouldn't make a bit of difference, for the following reasons.

1. Each spin is an independent trial
2. The same number of numbers remain on the wheel at each spin
3. The RTM doesn't tell you when a bet is "due to hit" or when it's even "likely" to catch up.

 

Trilobite

  • Mature Member
  • ***
  • Posts: 334
  • Thanked: 115 times
Re: RTM
« Reply #7 on: March 27, 2015, 01:12:35 AM »
Admittedly the three points you raise are hard to argue against.

All I said was RTM will help you reduce volatility. Betting the EC’s will help you reduce volatility too. I happen to think that playing a low variance game is better than playing a high variance game.

There is inherent and underlying ebb and flow to games of chance like roulette.

It is the ebb and flow that should be targeted, not individual bets.

Waiting, watching, selecting, and reverse selecting all linked to MM can help you squeeze a few percentage points out of your wagers, which can have a significant impact on a low variance game.
 

Real

  • Fighting the war on absurdity one foolish idea at a time.
  • Hero Member
  • ******
  • Posts: 1378
  • Thanked: 201 times
  • Gender: Male
Re: RTM
« Reply #8 on: March 27, 2015, 03:16:19 AM »
Variance is a double edged sword.

It's what enables you to win and lose big.

So why limit it?
« Last Edit: March 27, 2015, 09:54:05 AM by kav »
 

Trilobite

  • Mature Member
  • ***
  • Posts: 334
  • Thanked: 115 times
Re: RTM
« Reply #9 on: March 27, 2015, 04:44:49 AM »
You’re right.

All I know is most of the time you’re treading water and close to breaking even, which is fine. Then for a moment or two things start to go north or south. When it goes north you ride along, when it goes south you flip the poles and ride along.

RTM helps me do that. And most of the time you do that you're still treading water and close to break even, which is fine.
Like I said, a few percentage points and some clever MM is all you need in a low variance game.
« Last Edit: March 27, 2015, 09:54:26 AM by kav »
 

Real

  • Fighting the war on absurdity one foolish idea at a time.
  • Hero Member
  • ******
  • Posts: 1378
  • Thanked: 201 times
  • Gender: Male
Re: RTM
« Reply #10 on: March 27, 2015, 04:54:30 AM »
Actually RTM isn't doing anything for you.  It simply is, after the fact.  Meaning after it's happened, you can't exploit it.  Furthermore, you don't know when or if it's happening.  There's nothing there that you can exploit or use to help reduce draw downs/losses.  You have no way of knowing whether a streak will continue or end.

My suggestion is to stop trying to limit variance, and try for a larger win.
or
If really hate the variance, then bet every number so that there's no variance, just the house edge predictably removing chips from you at each spin.

Are you in Europe or the US?

Best of luck,
Real
« Last Edit: March 27, 2015, 09:53:44 AM by kav »
 

Sputnik

  • Mature Member
  • ***
  • Posts: 494
  • Thanked: 390 times
Re: RTM
« Reply #11 on: April 06, 2015, 04:04:17 PM »

Well Real is right again :-)

I test this and got the code:

 10 trails is 4.52 SD (20 in a row
 11 trail is 4.74 SD ( 22 in a row
 12 trails is 4.96 SD ( 24 in a row
 13 trails is 5.16 SD ( 26 in a row
 14 trails is 5.35 SD ( 28 in a row
 15 trails is 5.54 SD  ( 30 in a row

 This is when you take for example 10 random outcomes to match 10 future ones.

 Is not working,
 I run 2000 trails several times and when i up 20.000 trail almost all have hit 20, 22, 24, 26, 28

 So this is not better then betting agains't 10 11 12 13 14 15 in a row
 

kav

  • www.Roulette30.com
  • Administrator
  • Hero Member
  • *****
  • Posts: 1830
  • Thanked: 801 times
  • Gender: Male
Re: RTM
« Reply #12 on: April 07, 2015, 12:20:25 AM »
I run 2000 trails several times and when i up 20.000 trail almost all have hit 20, 22, 24, 26, 28
I don't understand what you say there.
Please explain in more detail.
 

Bayes

  • Veteran Member
  • ****
  • Posts: 688
  • Thanked: 550 times
  • roulettician.com
Re: RTM
« Reply #13 on: April 09, 2015, 11:25:37 AM »
All I said was RTM will help you reduce volatility. Betting the EC’s will help you reduce volatility too. I happen to think that playing a low variance game is better than playing a high variance game.

There is inherent and underlying ebb and flow to games of chance like roulette.

It is the ebb and flow that should be targeted, not individual bets.

Waiting, watching, selecting, and reverse selecting all linked to MM can help you squeeze a few percentage points out of your wagers, which can have a significant impact on a low variance game.

Yes! All you have to do is lower the variance sufficiently so that you can use a mild progression (no martingales!) in order to come out ahead. Aggressive progressions will kill you every time.

I agree it's hard to come up with the math or a decent argument to refute Real, but there are mathematically sound principles for reducing the variance (look up "Variance Reduction" on Wikipedia).

One of these is "Stratified Sampling". You divide the total "data" (outcomes from the wheel) into mutually exclusive and exhaustive categories, then make sure that you bet on each of these categories systematically. Anyone who has ever tried "mixing up" bet selections knows that this is an improvement on just using one exclusively.
 

Paulnewman

  • New
  • **
  • Posts: 23
Re: RTM
« Reply #14 on: April 09, 2015, 12:00:13 PM »
It is always the best way to adjust the betting according with the expectations, statistic and probability.

Every time you play a EC, your probability to win is around 48,7%(it is not perfectly 50% because the presence of the 0), and this data not depends on the strategy that i’m using for choose the black or red.
So every time i’m betting on black or red i have this probability to win.

In 10 spin, i should win around 4 times and lose the other 6.

So if you are winning more than 4 times, you need to expect some losses in the next spin.

If you are winning less, you can expect that things will run better. For example, if you won only 2 times and lost for 8 times, you can expect that in the next 10 spin you will win 6 time and lost 4.
It is easy and simple to adjust the progression with these data...