Hi Mark,
This "Roulette & gambling framework" subforum is the precise spot for this type of discussion!.
I like you added the "lifespan" dimention. This is in line with arteinvivo's explanation of trends, adding a time dimention to it.
Regarding me, my "time" dimension is a "cycle". These should be respected and performance measured in batches of their "ideal frequency".
For instance when tracking a 12-number event (May it be a dozen, or two lines or 3 corners... whatever the 12-numbers are), the cycle to count is is 3 spins.
So in the "timeline" of that 12-number event:
1 in 3 spins = Expected.
2/3 in 3 spins = More than expected.
0 times in 3 spins = Dispersion.
For a line it is 1 in 6 spins.
For a corner it should be 1 in 9 spins.
For a split 1 in 18, etc. etc.
A full cycle of spins gets calculated according to 36 divided by the amount of numbers covered.
==================
THIS IS WHEN IT GETS FUN
==================
What we call "trend" is an agrupation of numerical occurrences in a timeframe. We can track whatever we want and say: This is trending, this is coming as expected, this event is having dispersion.
In this regard, whatever you are tracking, the time to enter is when it is "trending" i.e. coming more than expected. You have to drop it when it faces dispersion.
The importance of having selection criteria is for us to determine what is "HOT" and "WARM". This is basically what we should be betting. Never cold locations.
The only exceptions for betting cold are the "Statistical return" and the "Statistical max" areas.
The "return" is the most common point where a sleeper comes back. They do form a bell curve. Say for a dozen at your casino's wheel it is most common for a sleeping dozen to come back from spins 9 to 12, then this is YOUR return zone.
The "max" is obviously the higest point of absense withnessed by the punter.
In my opinion think these values should better be held as a "per person" basis rather than in a generic "worldwide" form.
For instance some of us will die without seeing a streak of 30 reds or 30 blacks. BUT others will see it. Therefore betting the "statistical max" will be profitable for all those guys seeing 20+ but never reaching the 30's.
==================
Small excercise:
Go to
www.random.org and generate several sets of random lines for a location:
1 to 3 for dozens.
1 to 6 for lines.
1 to 9 for corners.
1 to 12 for streets.
1 to 18 for splits.
Pick one location for the excercise. Say you go for the "lines". Highlight the hits and jump from line to line when it concentrates hit. Remember:
============================================
NOT EVERY CONCENTRATION OF HITS TURNS INTO A TREND; BUT...
============================================
...But every trend starts with a concentration of hits.
Meaning for an excercise like the above your selection criterion is "random selection" and you are entering only after a concentration of hits (2 or more shows in a cycle) and hold your bet for as long as it is concentrating hits OR coming as expected: 1 time per cycle. Get out of the selection criterion as soon as it faces dispersion. I.e. not coming even 1 time at a given cycle.
------------------------
That's a fun excercise. I remember having done it for lines just to be certain the concentration/dispersion in Lw's is the same for random selection... and it is.
The basic concentration/dispersion-surfing techniques can be applied to that, but still I prefer a traditional criterion to the random line for real-money play :)
Have fun,
Your friend.
Victor.