Troubling Negative Variance

[Frostbyte]

I hit two stores a long distance away this weekend. At the first, I was up 700 in three hours playing all at $5 and $10 tables on a 10k bankroll at approximately half-Kelly. I would have preferred to back-count but there were never more than three tables open in this store and the floor, over the remainder of this and the following session, took notice of me. In the next two hours I lost this entire win plus an additional 600. The psychological impact of this was such that I ended my session to avoid going on tilt.

The next morning I returned to the same store and lost an additional 400 back-counting. I cashed out what I had and headed south to the second store, which had enough tables for me to be able to wong in and out freely. Playing 25-75 with a wong-in at TC+2 and wong-out at zero, and departing from a shoe that had not reached TC+1 at 50% dealt, betting 25*(TC-1), I lost my entire $400 buy-in in the space of a shoe, bought in again and as the TC soared to +7 and even to +10 at the cut card I ended up with a 150 loss for the evening. I won $275 wonging freely again the following morning and then dropped back down to my original 150 loss in the next shoe despite the RC peaking at +31 with 3.5 decks left, consistently getting stiffs against dealer tens.

My aggregate loss for some thirty hours of play this weekend is somewhat over a thousand dollars, roughly fifteen max bets or 12% of bankroll.

I believe these results are within the standard deviation but they are nonetheless troubling. Am I overreacting?

[sagefr0g]

what does half kelly mean? if it means about 6.8% ROR, that's a bit high of a risk of ruin, no?

edit: oh i guess you meant 0.5 kelly factor, no? not so bad as 6.8% ROR then i guess.
anyway, sorry to hear such a rough rocky road for yah. i'd be upset, but then i get upset easily.

[Frostbyte]

what does half kelly mean? if it means about 6.8% ROR, that's a bit high of a risk of ruin, no?

Unlike EV, the risk of ruin does not scale linearly with Kelly fraction. The probability of ruin on an infinite timeline at a given Kelly fraction is given by exp(-2/K) where K <= 1 is the Kelly fraction in use (Chin and Ingenoso, "Risk Formulas for Proportional Betting", http://www.bjmath.com/bjmath/proport/riskpaper1.pdf)

Consequently, betting at, say, one-eighth Kelly will give one eighth the geometric growth rate accessible through full Kelly at a miniscule fraction of the ROR (for K <= .2 the ROR is so small as to be measured in thousandths of a percent or less).

For half-Kelly the above formula gives an ROR of 1.832%. This is still higher than I would like under the circumstances, but necessitated by the piddling hourly EV my bankroll can offer.

[Frostbyte]

You are overreacting. 15 max bet swings are not that uncommon perhaps even common place. When you wong you are subject to bigger swings generally speaking. The advantage you have at higher counts is from the doubles, splits and blackjacks and how they are more prevalent and more profitable. If you don't get them you will win about the same percentage of hands as at any other count. If you get the profitable doubles and splits and lose them you lose even more than you would have otherwise. Then sometimes the planets align and you get nothing but doubles and splits with big bets out and win a lot of money. You say you played 10 hours or perhaps 1000 hands. Play 30 or 40 times (about what your N0 is) that and you should expect to not be in negative territory. You will have lots of trips like this one or worse. When it gets tough is when you get lots of trips in a row like this. If you can't play through these common place downturns without going on tilt you shouldn't be attempting to earn money at blackjack because you are not disciplined enough to succeed. Many have failed because of this.

Some separation from the events has redeemed my resolve. I know from the markets that a 12% swing is hardly irrecoverable, and I'll be playing again later this week and again next week. This is an unprecedented swing for me, and my first and only downswing so far, and I'm sure that familiarity will make it easier to take.

[Matt21]

Hi Frostbyte, sorry to hear that you have had this bad swing.
When I was counting, I always found it difficult to handle losses that were greater than any loss I had experienced previously. Each time I would hit a new loss by a wide margin, it would hit me psychologically and it would take me some time to mentally recover. I would run through the numbers to see how likely/unlikely that really was, and then consider whether I had done something wrong or whether the casino had cheated me in some way - but of course each time it was pure variance. So I am not surprised if you are asking this kind of question after suffering a big loss like this for the first time. There will be more and bigger ones in the future - but likewise with the winning sessions.
Good luck out there.

[Anton Chigurh]

15 maximum bets does not constitute an unusual loss or win over a few hours.

There will be many such downswings in your future. May they be widely dispersed amidst many hours of less traumatic play.

Large wins or losses on Wong-ins are not unusual for me: in at +2, kiss off 12 maximum bets as the dealer pulls 20 and 21 over and over, and then leave at 0. Or watch the count go to +infinity in the first two rounds, place 6 max bets, push 1 and lose 5 before Wong-out.

I often wonder if I pissed off someone in a past life. I know players who have played years longer than I have and not seen some of the downswings I have seen. Remember: even the variance has variance.

[Aslan]

My aggregate loss for some thirty hours of play this weekend is somewhat over a thousand dollars, roughly fifteen max bets or 12% of bankroll.

The only thing you can count on in card counting (in the long run) is that slow accretion due to your expected win rate, and that is only if you play close to perfectly. This is why I recommend that when you have those great upswings that exceed your win rate, you do not spend them. They are not yours (technically they are) because if and when the large downswing in variance occurs, you will have to pay the piper. They are yours, but my experience tells me that the roller coaster ride in card counting is the norm, and therefore you can count on the downswings, some of them large, like you can count on the sun coming up each morning. They're not predictable, but they do come sooner or later.

If your were wise and did not spend the difference between your expected win rate and the peak of your win amount, you would be able to fade (that is, tolerate) the downswings a lot better. If you did not do that, look at your records if you keep them, or think back if not, and try to remember how much you may have exceeded your expected win rate during the preceding year. This will make your loss feel a lot more palatable, and speed up the recovery of your depressed mindset. Just my two cents. YMMV.

[Most Interesting Man]

Been dropping a minimum of $1000 almost every day for the last few weeks. Went from being in the lead with 1.5x my starting amount down to two thirds. These things happen. In the end it will turn around.

[White Guy]

Funny I just posted about how pissed I was about a massive dump I took on two consecutive shoes earlier.

Here is an example of my last 40 hours of playing that is not uncommon. I was playing HU mostly decent to good games spreading 1-24 when I can but at least 1-16.

8 hr period -80 units
8 hr period +80 units
8 hr period -320 units
8 hr period +440 units
8 hr period -170 units

I rounded the numbers a little but This is not at all uncommon.

Counting from my very lowest to my highest point I had about a 500 unit swing. Almost all of that swing was in two really really bad shoes with a 12 unit max bet followed by 2 Good shoes with a 24 unit max bet. Oddly the two bad and 2 good were both consecutive within about 8 hrs from each other..

If I were betting the same max bet the loss would have been twice as bad as the win so it is just dumb luck that it happened like that.

Total played probably 5-6k hands.

That is why I think it is insane to play with less than a couple hundred max bets.

I play with 0 ROR and it still pisses me off. Just today I had a dealer make 3 21s in a row with 24 unit max bets out. NO ****..

I couldn't imagine getting my ass handed to me knowing there is a chance I could actually ruin myself. To me that is insane.

The EV will show but you don't know when.

[Aslan]

8 hr period -80 units
8 hr period +80 units
8 hr period -320 units
8 hr period +440 units
8 hr period -170 units

As if that is not enough, what really tests one's mettle is when the three negative 8-hour periods are consecutive! My neighbor says, "Old age is not for sissies!" Well, card counting is not for sissies either.

[Coyote]

Frostbyte, take solace that so many of us have been in your exact situation. A little time away from the tables helped me recharge mentally also. What I am finding, as the years go by, is that the variance gets a little easier to handle....a little easier. During my short break, I found where I could tweak my game a little to lessen the impact of those down swings.

No matter what, Have Fun! We will all be gone soon enough!

Best of cards to you all

[Goatlife]

[QUOTE=White Guy;20254]


8 hr period +440 units
QUOTE=White Guy;20254]

how do you not get skills checked when you have a massive win like that ( I am assuming that 8 hour win games from some heaft wins in a single shoe and a single casino

[Frostbyte]

[QUOTE=smallcapgrowth;20285]

8 hr period +440 units
QUOTE=White Guy;20254]

how do you not get skills checked when you have a massive win like that ( I am assuming that 8 hour win games from some heaft wins in a single shoe and a single casino

I would think the house is just as accustomed to variance as we are. Someone watching him might look at forty hours and conclude he's a standard loser.