Kudos to Tthree for his last post.
The formatting of a table here is a problem,
at least without using the "advanced" window
For those who do not know, the D.I. is the Desirability Index
and that is roughly the square root of the SCORE.
Kudos to Tthree for his last post.
The formatting of a table here is a problem,
at least without using the "advanced" window
For those who do not know, the D.I. is the Desirability Index
and that is roughly the square root of the SCORE.
Last edited by ZenMaster_Flash; 08-22-2016 at 10:56 AM.
I don't know why those on this site have such a desire to make this more difficult than necessary. It seems like some just have the need for extra mental gymnastics, like they are trying to prove something to the world.
I have said this many times when saying that how pros view gain will hurt small BR players. Everyone poo-pooed me. I am glad someone understands what I was saying.
It is not gains/RoR it is SD/EV. I showed how what is likely too follow any given outcome is affected by this stat. I have made tweaks to games other than BJ that more than double EV with a modest increase in monetary SD. You need to go 16 SD's in the negative direction before the frequency at the same monetary loss becomes higher for the improved count. I will give a brief summation to show the effect:
Original; Tweaked
EV: $933.97/1200 rounds; $1930.24/1200 rounds
SD: $4969.73/1200 rounds; $5032.87/1200 rounds
Ratio EV/SD: 5.32/1200 rounds; 2.61/1200 rounds
Now look at the monetary EV per for each number of SD's when 1200 rounds have been played:
-3 SD's: -$13,975.22; -$13,168.37
-2 SD's: -$9,005.49 ; -$8,135.50
-1 SD's: -$4,035.76 ; -$3102.63
Expected : $933.97 ; $1930.24
+1 SD's: $5,903.70 ; $6,963.11
+2 SD's: $10,873.43; $11,995.98
+3 SD's: $15,843.16; $17,028.85
As you can see, assuming the same frequency at each SD, you are far higher EV at every similar frequency. Like I said you would need to go to the -16 SD point for the original to lose less money at the same frequency (# of SD's). If you wanted to equate dollars rather than SD this is what it looks like in more detail (these frequencies are for actual sim results which do not assume a standard bell curve. The original was left skewed bell curve and the tweaked had a right skewed bell curve. Skew was relatively minor at original skew -0.0017 and tweaked skew +0.0357. this minor skewing adds a hair to the shift in frequency of likely results at the extremes):
Frequency; (range of results) Technique
-------0.1847%; (-$13975.21 to -$12732.79) Original
0.1697%; (-$13168.36 to -$11910.15) Tweaked
-------0.3957%; (-$12732.78 to -$11490.36) Original
0.3194%; (-$11910.14 to -$10629.44) Tweaked
-------0.6053%; (-$11490.35 to -$10247.92) Original
0.5894%; (-$10629.43 to -$9373.47) Tweaked
-------1.0357%; (-$10247.91 to -$9005.49) Original
1.0139%; (-$9373.46 to -$8135.50) Tweaked
-------1.6599%; (-$9005.48 to -$7763.06) Original
1.6667%; (-$8135.49 to -$6877.28) Tweaked
-------2.5829%; (-$7763.05 to -$6520.63) Original
2.5775%; (-$6877.27 to -$5619.07) Tweaked
-------3.7485%; (-$6520.62 to -$5278.19) Original
3.7938%; (-$5619.06 to -$4360.85) Tweaked
-------5.2092%; (-$5278.18 to -$4035.76) Original
5.2709%; (-$4360.84 to -$3102.64) Tweaked
-------6.7258%; (-$4035.75 to -$2793.33) Original
6.8341%; (-$3102.63 to -$1844.41) Tweaked
-------8.2447%; (-$2793.33 to -$1550.90) Original
8.2892%; (-$1844.40 to -$586.20) Tweaked
-------9.4058%; (-1550.90 to -$308.46) Original
9.4824%; (-$586.19 to $672.02) Tweaked
-------10.0782%; (-$308.45 to $933.97) Original
10.1174%; ($672.03 to $1930.24) Tweaked
-------10.0188%; ($933.98 to $2176.40) Original
9.9413%; ($1930.25 to $3188.46) Tweaked
-------8.2202%; ($2176.41 to $3418.84) Original
8.1231%; ($3188.47 to $4446.68) Tweaked
-------6.7468%; ($3418.85 to $4661.27) Original
6.6857%; ($4446.69 to $5704.89) Tweaked
-------5.2062%; ($4661.28 to $5903.70) Original
5.1337%; ($5704.90 to $6963.11) Tweaked
-------3.7801%; ($5903.71 to $7146.13) Original
3.7547%; ($6963.12 to $8221.33) Tweaked
-------2.5808%; ($7146.14 to $8388.57) Original
2.5737%; ($8221.34 to $9479.55) Tweaked
-------1.6668%; ($8388.58 to $9631.00) Original
1.6869%; ($9479.56 to $10737.76) Tweaked
-------1.0325%; ($9631.01 to $10873.43) Original
1.0436%; ($10737.77 to $11995.98) Tweaked
--------0.6047%; ($10873.44 to $12115.86) Original
0.6303%; ($11995.99 to $13254.20) Tweaked
--------0.3428%; ($12115.87 to $13358.30) Original
0.3563%; ($13254.21 to $14512.42) Tweaked
--------0.1829%; ($13358.31 to $14600.73) Original
0.1983%; ($14512.43 to $15770.63) Tweaked
As you can see in the positive results the tweaked range is much more similar to the Original listed below buts its frequency is more similar to the Original that made less. In negative territory the same is try but in that case losses less more frequently instead of winning more more frequently. It is easy to see that losses are less likely to clump and wins are more likely to surround losses and that the magnitude of each likely event favors the player winning by a lot. winning smaller amounts is over 50% more likely than losing the same amount and winning large amounts is twice as likely as losing small amounts. Without the difference in ratio of SD/EV the gains of the higher EV count would show more frequent losses (or bigger losses at the same frequency) before you get to 2 SD's negative. There is no swing control there.
It isn't about lowering variance. This isn't affecting variance. It is about controlling swings and the math is quite state forward. First you need to increase EV. I think everyone wants that. Then if the monetary SD increase is low compared to the monetary EV increase swing control starts to be possible. I just demonstrated how it works. You are changing the odds that you sill get large losses and changing the odds of all the outcomes so that what is likely to surround it quickly erases the loss. It is simple odds of various win and loss magnitudes interacting to show how things are likely to unfold over larger time intervals. Results follow the expected combinatorial analysis of these odds for the various ranges. the result is you make more money over time and you sty much closer to your expected higher win rate than with the lower win rate. If you want to make less and have results that are not very predictable that is up to you. I prefer predictable results that make more money in the same amount of time.
Swing control? Again, a psychological flaw in the context of advantage play to believe that that's important. Repeating the same thing 11,000 times until anyone else either gives up or blows up and is banished doesn't make you any more or less correct, nor does the tactic of using big words and fonts.
Last edited by mcallister3200; 08-22-2016 at 12:15 PM.
Like I said I prefer to make more money with greater certainty of outcome. You like to make less with a lot less certainty that you will continue to grow your BR and much larger swings while getting the pay cut. If that is the unflawed pro approach I will take the flawed one. My approaches and techniques to find solutions have taken an extremely wild BR swing game and made it into a game with very moderate BR swings and quite certain wins in a time frame that would have most dot sure whether they would win or lose big or anything in between. For BJ things aren't quite as easy to alter. The tweaks bring more modest gains so time frames to notice the change in outcome are much longer. I understand the concept that you have a million dollar BR so you don't care if you lose a few hundred thousand. I still would care. I would have a higher EV than you and not have to endure such swings because we can both make that choice and I would choose differently. Like I said I will take the raise in EV and the taming of swings. I showed you the math and mechanism of longterm swing control as well as short term swing control. I know after 1,000 hours I want to be very close to EV not a good chance to be way below it or even in the red. I think any true professional wouldn't be able to argue with that. If they do they are just making themselves sound like rational idiots trying to justify something rather than make a good business decision.
do you always speak for other people and put words in their mouth or just on the Internet? You are clearly trolling at this point. I never said anything of the sort. If you want to speak for me and say I would rather expend my energy searching for much larger unrelated edges then that would be accurate instead, but do not speak for me.
"Women and cats will do as they please, and Men and dogs should just relax and get used to the idea" --- Robert A. Heinlein
That just sounded like what you are arguing. Swings make a difference. The way EV grows linearly and SD grows by inverse exponent, square root of rounds played. Basically n0 or SCORE will be factored in to what I have shown controls swings as well as the range you will be likely to be away from EV after a large number of hands. Basically poo-pooing what I have been saying is poo-pooing increases in SCORE. I just go into more detail. At a certain ratio of SD:EV you are increasing SCORE or long term results range and if the ratio is changed large enough you see short term swing control. It is the mechanics of why you get a better SCORE and why weaker counts tend to have poor results stack up more frequently and to a higher decree than stronger counts. I showed the examples that show how the interaction between SD and EV cause these thing to happen. Don calculated a version of this and called it SCORE and we have n0. These things are not overrated I just got further into how and why they affect results than most do and you didn't seem to care and started to act like SCORE and n0 are not a concern. You can get more specific than that as I have shown.
Last edited by Three; 08-22-2016 at 01:20 PM.
It is funny. On another site they say I was a red chipper a few years ago and now I am against casino tolerance limits and somehow they decide this proves I am a fraud. The idiots don't even consider the fact that if I am right this is exactly what you would expect. I would quickly grow a BR from red chip to not being able to increase my bets because I reached resizing amounts quicker and with a higher degree of certainty. You are right though once against the tolerances much of the gain is lowering my RoR which is already below .01% RoR. While I don't mind the increased certainty of steady winnings there is not a reason to think it makes my BR any significant amount safer. I am on an insane winning run right now. When I complain about my bad stretches it is I only won $1K in a month not that I lost $20K in a month. In the end the pro with the $20K swings has about the same EV as I do. The difference is at the end of the year I will almost certainly be very close to EV and he will have a wide range of expected results that even include a losing year. That means I am unlikely to be at double the EV where he may very well be but I like certainty of income. It is much less stressful.
Last edited by Three; 08-22-2016 at 01:23 PM.
Bookmarks