Hi there! I'm new here, so here's a very quick introduction: I'm
Jean-Baptiste M. Queru, French male 30-something living in the San
Francisco area, with a background in mathematics and software (I'm a
software architect by trade). I'm interested gambling strategies that
are both mathematically sound and practical. Yes, Jean-Baptiste
(hyphenated) is my full first name, and yes it's a guy's name in
French - but you can call me JBQ.
I'd like to discuss and possibly challenge the wisdom of always
playing "max coin" in VP. I'll focus on JoB as it's the game that I
personally like, but most will apply to other games.
My ski trips to Reno exposed me to a lot of 9/6 JoB machines in all
denominations (to quote the vpFree data, Atlantis has denominations
all the way from 1c to $100). I found some strategies on the web,
trained a bit, and used my next opportunity in Reno to spend some time
playing on penny machines, which was lots of fun.
In a recent trip to Vegas (for which I was admittedly unprepared), I
was disappointed to not find much 9/6 JoB in low denominations (I know
that there are some, but you don't find them by chance walking on the
casino floors of Strip casinos between Flamingo ave and Tropicana
ave). I ended up playing a bit of 8/5 JoB but some health and sleep
problems somewhat ruined the trip so I didn't really miss 9/6 that
much.
So, as I wrote above, a number of factors made me consider whether the
common wisdom "always play the maximum coins even if you need to drop
to a lower coinage":
-Playing maximum coins creates a game with a very high volatility. As
an example, 5-coin 9/6 JoB has a variance of about 19.5, meaning that
a 5000-game day has a SD of over 312 (and we all know that 5000 games
really isn't much and that it's easy to be several SDs away from the
EV. The ER by comparison is 4977, i.e. an expected loss of 23.
Personally, the SD drives me away far more than the expected loss. By
comparison, playing 1-coin has a variance of about 5, meaning that the
same 5000-game day will have only have the SD - 156. The ER falls to
4918. I might feel more comfortable with a game that is worth -82 +/-
156 than with one that's worth -23 +/- 312, the extra 59 coins buy me
some peace of mind.
-Sometimes (and seemingly very often on the strip), dropping to a
lower coinage might mean dropping to a less favorable pay table.
Playing 1-coin 9/6 JoB has only a very slightly worse return than
5-coin 9/5 JoB or 5-coin 8/6 JoB, and is significantly better than
5-coin 8/5 JoB which is commonly found in many areas. Playing 1-coin
$1 9/6 is a far better idea than playing 5-coin 25c 8/5.
-Playing a higher denomination can be the key to the high-limit area,
where cocktails and other such floor comps are typically better, and
where you have typically more space, more attention from hosts and
less traffic.
-In the case of 9/6 JoB at least, the optimal strategy for 1-coin
seems to be simpler than for 5-coin (i.e. there seem to be fewer cases
of penalty cards), though in all honesty I haven't tried yet to
memorize the details of the discussion about 3 to a flush with 2 high
cards vs. only 2 high cards. It falls into 3 distinct patterms and so
shouldn't be that hard, and in my opinion easier than the discussion
about suited high card and T in 5-coin.
-Or, simply (at least in 9/6 JoB), the expected absolute loss of a
single game at 1-coin is lower than than that of 5-coin. If you're
already playing at the lowest denomination and want to make your
bankroll last, playing a single coin at a time is certainly worth
considering (after all, in absolute terms, 5-coin loses money 40%
faster than 1-coin and has a SD about 10 times bigger)
As a conclusion, I will say that I can find several reasons why
playing fewer coins than the maximum can be worth looking at. Some can
be proven to be mathematically better, some can't quite be quantified
as easily.
JBQ
PS: For those who are interested, I can post my strategy for 1-coin
9/6 JoB. I originally thought that such a strategy would be easy to
find online, but since all the research seems to focus on 5-coin I
didn't actually find it so I decided to research it. I didn't actually
prove my results to be correct, but so far they have been tested
accurate with about 360000 random deals.
