2018 9/6 job trial

[olds442jetaway]

   In hopes of not repeating the 2017 disaster results, I have started a new trial for this year. Knowing in advance that job will be a failure unless the Royals and quads come in on time, I started this experiment. Luckily, I have 1 machine at Mohegan sun that pays 800 for 1 for bet 1, so I will be sticking to this machine. So far, I have just had 1 trial. My second trial will be tonight. Starting with bet 1, anytime my bankroll exceeds what I started with, my bet will remain at bet 1. When I have a net loss of 25 coins at bet 1, my bet will go to bet 2. Anytime playing bet 2 my net loss is less than the original 25 coins I lost at bet 1, I will return to bet 1. For example, in my first trial I had a net loss of 25 coins after an hour or so. I started betting bet 2 and several times when I hit my net loss was under 25 coins, I returned to bet 1. Lets say that I returned to bet 1 because after hitting at bet 2, I only had a net loss of 18 coins. I would remain at bet 1 until my net loss was 25 coins again. Well, of course that happened and I was forced to bet 2 again. At bet 2 after I lost 50 coins eventually and was forced to go to bet 3. I just started at bet 3 when it was time to go home. When I go back tonight, I will be at bet 3 unless I lose 75 coins at bet 3 or hit something where i could return to bet 2 or even bet 1. The hope of course is using this very very mild martingale way of playing that I will hit quads or royals at the higher bets. If my bets get too high where I will lose my entire bankroll, of course I will end the trial and declare this method a failure. Anyway, we will see what happens. Hope for a quad at bet 3 when I begin tonight. So far this has kept me playing a long time with minimal money. My bankroll will be 500 bucks. So far, I am only down about 80 bucks and I am playing a dollar machine. If I hit the Royal at bet 3, I will win 2400 bucks. I prefer to play this way on quarters, I may switch to quarters if i get wiped out at the dollar machine.

[Gronbog]

Betting more than minimum based on any criteria which does not involved identifying a positive edge will only increase the rate at which you lose your bankroll.

[billryan]

   Recreational players don't care about mundane things like profit. They simply want to have fun.
Math is irrelevant. It's the high they are chasing.

[Gronbog]

I agree and that's not a problem. To each his own. It just sounded like the OP thought that this might make his bankroll last longer.

[New2vp]

Hey guys,  while Gronbog and billryan are not wrong, there is more to this than that.  The math is considerably more difficult than simply picking the maximum EV play that is necessary with video poker math.Also, this needs to be compared to something.  9/6 Jacks is often a "good enough" game without specifically defining the quoted term.Since the game that olds is suggesting has no royal penalty for single-coin betting, there is no house-edge difference between betting one, two, three coins, and whatever max coin on the machine is.  We also don't know at what level olds would choose to bet if he were flat betting instead.  So without working out the math, it is a little harder to figure out what olds average bet size is based on what he has described.  That is not undoable, but would take more time than it takes to post this.  Also, olds has not described precisely how he chooses to stop a session.  It could be based on being so many coins ahead or after so many hours of play or some combination of the two.  Of course, we know about the stopping on the negative side, when his daily bankroll is depleted, but to work out the details, we would need to know the positive stop strategy as well.The reason that people gravitate to martingales, mild or not, is an attempt to change the outcomes.  And regardless of what flat-bet enthusiasts would say to that, the odds do change.  Is there rationale?  For those who are simply attempting to maximize expected $ gain or minimize expected $ loss, there is really no difference in expected value at martingale betting vs. flat betting if the flat betting is the same as the average bet size of the martingale betting.  Of course, there is a difference in variance.But the odds of having a winning session can go up depending on how the martingale is designed.  Please don't accuse me of math heresy yet.  If this doesn't make sense to you, stay tuned.Consider a 4-hour, 600-hand-per-hour session of flat betting 9/6 Jacks or Better.  This will give you a 30.35% chance of having a winning session, a 69.34% chance at a losing session, with the leftover small balance being an exact push.  For those who really hate losing, that doesn't sit well.  And how does a 99.54% game produce such uneven results?  Well, the average win is much bigger than the average loss.  With the experiment that I defined above, the average win is 227% of the size of the average loss.  Reciprocally, this means that the average loss is 44% of the size of the average win.If you simply count winning days vs. losing days, the losses way outnumber the wins.  Some people "feel" way worse with a loss than the amount that they "feel" good with a win.  It isn't "linear," without trying to define that term at this point. So what can we do about this?  Well, you may not want to spend another 4 hours flat-betting.So you can decide to start out betting less and then bet more if you turn out being behind.  There are definitely many martingales that can be designed to change the odds of an overall win to 50/50 vs. a loss.  Of course with Jacks or Better, that is going to change the balance of expected win vs. expected loss.  With any 50/50 martingale, the average losses will increase to a point that they are a bit greater than the average win (about 101%).Well, some are not as unhappy with the less favorable win amount to loss amount ratio as they are happy with the more favorable ratio of win frequency to loss frequency.We can similarly design martingales to win 75% of the time, 90% of the time, or even 99% of the time.  The problem comes up because human's pattern recognition is not based on the math, but is often based on relatively recent experiences.  With all that winning going on, you may believe that you have happened onto a system that seems to perform better than it does.  It may take a long time with a 90% pattern to experience a loss.Let's look to see what happens when the losses do come.  With a 75% system, the losses average 306% of the average win size.  With a 90% system, the losses average about 9 and a half times the size of the average win.  And with a 99% system, the losses are 182 times the size of the average win.  In the situation olds described, if losses are really bounded by the $500 daily bankroll, that will require the average win size to go down or the frequency of max $500 losses to go up.  Of course, for those who really, really hate to end the day with a loss, there is always a temptation to go beyond the amount that they previously set as a limit if there is some possibility of getting one's hands on a little more than the original $500 limit.I cannot tell exactly what percentage system that olds has designed here both because the stopping point is not fully defined and I haven't set aside the time to analyze it.The problem with any martingale occurs if one is unprepared for the magnitude of the losses when they come.  Now, we can get back to talking about what I mean by linear.  If you have gone 9 days in a row with winning $50 per day, it is easy how to forget how unhappy it makes you on the 10th day when you give it all back and then some.  And it's even less fun when those rare occasions occur when the losses come in unexpected bunches, more closely together than expected.  All these systems work well until they don't.  There is nothing intrinsically wrong with them UNLESS you believe that you have found a system for overcoming the house edge and you are unprepared for the consequences when they occur.Olds, I'm not picking on you.  I wish you nothing but good luck.  Here's hoping for better than expected results.  That is certainly possible.  And only you can figure out what system maximizes your own happiness function, even if that math is harder than standard video poker math.

[Gronbog]

If percentage of winning sessions is the goal then, I agree. Progressions will help to accomplish that goal. That's why they are such a powerful tool in tournament play (almost any game). I just didn't want him to think this was a way to keep his head above the water long term.

[FAA]

New, assuming that I got woke and commit to quarter 9/6 JOB, what is the prognosis after 1,000 hands in two hours, my expected session? Great post.

[New2vp]

New, assuming that I got woke and commit to quarter 9/6 JOB, what is the prognosis after 1,000 hands in two hours, my expected session? Great post.Thanks, FAA.  Here are the numbers. courtesy of VPFW.  If you are terribly unfortunate with a $250 bankroll, you could run out of money around hand 762 about 1 time in 10,000 and around hand 994 about 1 time in 1000.Overall with that bankroll, using max-EV strategy and betting 5 quarters at a time, you are expected to win 34.57% of the time, lose 64.87% of the time, with an exact push the remaining 0.56% of the time.The distribution looks like (with the red numbers being losses):



Gain Between
  Frequency



$250.00
0.10%


$248.75
$203.75
0.58%


$202.50
$156.25
3.44%


$155.00
$110.00
10.32%


$108.75
$62.50
19.85%


$61.25
$15.00
24.09%


$13.75
$31.25
19.46%


$32.50
$78.75
11.95%


$80.00
$125.00
5.22%


$126.25
$172.50
1.88%


$173.75
$220.00
0.52%


$221.25
$266.25
0.12%


$267.50
$313.75
0.02%







$785.00
$830.00
0.06%


$831.25
$877.50
0.21%


$878.75
$925.00
0.44%


$926.25
$971.25
0.58%


$972.50
$1,018.75
0.54%


$1,020.00
$1,065.00
0.34%


$1,066.25
$1,112.50
0.17%


$1,113.75
$1,160.00
0.06%


$1,161.25
$1,206.25
0.02%







with the other 0.03% of the time either just over $313.75, just under $785, or if you're lucky enough to hit a couple royals or a royal with a bunch of quads or a couple straight flushes, over $1206.25.  Obviously, to reach the 2nd part of the table, it is almost certainly necessary to have hit a royal in those 1000 hands.

[FAA]

Interesting results. I'm often in the -$30+ range so quickly that I simply abandon ship for dollars, to exponentially worse results.

[billryan]

   That Bell Curve is for exactly 1,000 hands.
I think real life results will result in you breaking even much more than this indicates only because its a physiological sticking point. If you broke badly and played catchup all session, quitting at even feels like VICTORY.
If you break good, but eventually give it all back, quitting at even is an okay session.