Acceptable error rate

[Three]

Then are you saying your sims are inaccurate and that is why you overachieve EV?

I am saying I do things with extra information that the sim hasn't been instructed to do and it is based on many things including casino conditions which I would say can't be simmed at least simmed accurately. This means with perfect play I am expected to beat the sim. I have found it is not reasonable to expect a sim to properly reflect casino play. You can make a bunch of assumptions for the sim to make and hope it approximates it but it generally won't and the attempt introduces error to the sim. It is better to sim what can be simmed accurately and just know that if you are doing your work in the casino to get the best bang for your buck you will come in above that. Setting up wonging in 2d would not be that hard but simulating casino conditions for finding available games wouldn't be accurate. It is a bit more complex than linear wonging depending upon the 2 counts used and the game being attacked.

Now, if you are doing some sophisticated wong in out at various levels of TC above or below the sims thresholds then the sim may not represent your game.

The wonging applications in 2d with 2 counts will vary by the game you are attacking but it is very complicated as the system shows nonlinear relationships and a small change in cards removed at the same bet size may have little hope of changing advantage or a large change in advantage. In theory it can all be programmed but the availability of other tables, table crowding and other factors you deal with on the fly in the casino environment would just add errors into the sim if you make assumptions on what they would be. The one guarantee is the assumptions would usually be wrong. I prefer an exact sim that I know will have to be adjusted by varying conditions. I use ways to approximate what the ideal and impossible accurate sim would be but I usually just happy to come in above the sim results that have no assumptions to introduce error to the results versus the actual conditions you face.

I know statements like this will drive programmers crazy but I believe they are accurate. Programmers like to believe any conditions can be simmed but they can sim assumptions but assumptions are almost always flawed. Is it better to have a sim result you know is flawed and may or may not come close to the conditions you actually face. Or is it better to have an exact sim and understand the conditions you will face and what you do are not exactly what was simmed and your results with perfect play will reflect the difference.

Here is what I mean. You run a sim for a certain number of players or players bouncing in and out, a certain pen, perhaps include game speed assumptions, you make assumptions about table availability when you wong out, you set a wonging strategy that will not be able to process or reflect all the information, the count and beyond, that affect the efficacy of wonging, etc. The list could go on and on. Then you go to the casino and find varying pen, varying crowding, game speed issues, heat issues that need to be factored in, tolerance issues that need to be factored in, etc. The list could go on and on. I like a conservative sim so if I decide a game is worth it I will come in at or over the sim results. Given the differences between sim and actual conditions and the more assumptions you make in the sim the further you could get from the actual. I would suggest that coming in under sim results by 20% or more probably reflects the inaccuracy of the sim assumptions more than indicating you are making a boatload of errors. Perhaps those that do not seek perfection just make lots of errors. Like others I see the world as a reflection of myself and generally think few errors are made by AP's. I guess the fact that some AP's are so certain that high error rates occur is just an indication of what they experience themselves from their own play more than what others do.

[Stealth]

I would suggest that coming in under sim results by 20% or more probably reflects the inaccuracy of the sim assumptions more than indicating you are making a boatload of errors. Perhaps those that do not seek perfection just make lots of errors. Like others I see the world as a reflection of myself and generally think few errors are made by AP's. I guess the fact that some AP's are so certain that high error rates occur is just an indication of what they experience themselves from their own play more than what others do.

I have been a programmer and did work with some simulation systems that are many times more complex than you need in blackjack. So, no I do not have a fear of the structure being sim'ed if it can be defined. I have a problem with statements of certainty that are in an environment that has not been defined.

I would disagree with your statement that the error rate reflects the inaccuracy of the sim, at least in the context of my opinion and experience. I would suggest you are placing too much emphasis on the total cause of the variance from EV being due to playing/betting errors. There are a lot of items, many of which you mentioned, that contribute to the variance. And some of the issues you mentioned do not apply, in my opinion.

However, be cautious in you self analysis as I contend that you do not really know how many errors you are making. You only know about the ones you recognized. The very methods you are using to compare your results may be a large part of the issue.

A lesson I learned in a past life. I had occasion to work in a bank proofing department and had to balance at the end of each day. After about a years experience, I came to fear being out of balance a simple number like $10, when I often found that it did not represent one error of $10 but multiple errors of wildly varying amounts all totaling a discrepancy of $10. This concept is also very real in this game as blackjack and the nuances of AP play can do much to obfuscate reality.

In the end, I have no idea if your results have enough sampling to be meaningful, but if you are happy with your results and you think they are above an EV that is reasonably accurate, then good job.

Did the "translate steady profits" question get overlooked or do you chose not to share?

[Three]

Can you translate "steady profits" to a number like win 83% of sessions played on a basis of 345 sessions..?

Well last year I my records were kept a little differently. Entries were made based on a day's result's in a casino. if I added or took from my main wad multiple daily entries may have occurred. I tracked things between entries in my records by organizing where money was kept. I carried what used to be enough for a session as most people define it and it just rarely needs to be replenished in the day because profits are so steady. Sometimes my wallet would get ridiculously fat and I would do a dump in the middle of the day. Many of my losses are trivial just to book a loss and most of my wins are significant.

Anyway going by this method my entries last year went like this:
116 winning days/casino and 62 losing days/casino for 178 casino-day entries. About 65% winning casino-day entries.

If you throw out trivial entries both winning and losing:
115 winners and 36 losers for 76% winning casino-days

Now if you combine the to actual dates as opposed to casino with dates or multiple entries for the same casino-day:
83 winning days and 24 losing days for 78% winning days.

To put those daily numbers into perspective:
1 wins $5K to $5.99K
3 wins $4 to $4.99K
6 wins $3K to $3.99K, 2 losses $3K to $4K
12 wins $2K to $2.99K, 6 losses $2K to $2.99K
32 wins $1K to 1.99K, 3 losses $1K to $1.99K
29 wins less than $1K, 13 losses less than $1K

You can decide for yourself if that is steady winning but that is what I call it.

[Bodarc]

Pretty impressive numbers Tthree. Do you keep your sessions short or do you play marathon sessions? About how many hours of actual playing time would you average per day?

[Three]

Pretty impressive numbers Tthree. Do you keep your sessions short or do you play marathon sessions? About how many hours of actual playing time would you average per day?

It depends. Session length varies. I am trying not to win too much at a time. The results gives an idea of what that number is. I think that is about 350 hours of play if memory serves. If I hit my win tolerance goal I leave. If I show my supermax bet I take a break. If the deck compositions are too vanilla making my bet moves the same as others I play a shorter session. I probably average 1 to 3 hour sessions but lots are much shorter and under the right conditions I play longer. The numbers are not sessions the conventional way people talk about sessions but I am trying to answer you using session as the win before I take a break from playing at the casino or leave for the day. If I play at a casino cluster area I may get in 6 to 10 hours a day but if I am making a long circuit of casinos time played would likely be a little less. Or just playing one casino I might hit the tolerance threshold in a shoe or 2 if I am lucky. On a bad day I might spin my wheels all day and post a small win or loss because the count never did much. When I get killed in a day and post a decent loss either things are just going bad on big counts or I max split with doubles and have all my bets swept by the dealer. Fortunately the latter is fairly rare.

My 2 count application is much stronger than linear approaches for wonging. When betting EoRs get squishy advantage can change fast on the removal of a few cards or not have much hope to change with more cards removed. I really don't know the current betting EoRs but do know the effect certain card combinations removed would have on advantage as the 2 counts change. This knowledge implies the interaction of the squishy EoRs of the card combinations in question. You don't see this with vanilla deck compositions (evenly distributed ranks within card groups) for the same bet size which would make me easier to catch and cause me to play a shorter session. The unusual deck compositions that are not vanilla counts are where my correct bet moves can make me look like an idiot. I try to find and play these conditions more by my wonging style but I am not sure how effective you can be not knowing what the future holds. I know I have an effect but am not sure how much. Generally the higher the sum of the absolute value of each of the 2 counts the less vanilla the deck composition is but my decisions are based on the potential for actual shifts in advantage not the sum of the absolute value of the 2 counts. The tag overlap between the counts tend to push them in opposite directions but the overlap cards are balanced by different card groups so the balancing card groups removal can alter this general trend.

[Stealth]

115 winners and 36 losers for 76% winning casino-days

T3, thanks for the sharing.

When I read you first post on this matter it was not clear what level you were suggesting you could win. Our team results of over 3500 sessions (defined as any sitting at a table, 60,000 rounds) was to win 68% of the time. My partner explained to me before we ever created the team that his experience of major team play over the past 15 years provided a "rule of thumb" that you win at about 65% of your sessions. A rather loose concept that we later found to be accurate.

So, your experience of 76% winning over 105 days (~400 hours?? table time, 40,000 rounds??) is well within reason if you are playing with an advantage.

While I commend you on your results, I am not certain they represent any different results than produced by other count systems. I would dare suggest that you played a lot more negative counts in your play than our team did in theirs. We got our advantage by only playing in positive counts (BP play), you get yours by a stronger count (?). Bottom line is there are many ways to achieve the objective and the tool (count) used is only a small part of the answer.

What I call your steady winning is "normal".

BTW, how do the results compare to EV?

I am a big fan of anyone who is beating the system, so here's wishing you continued success!

[Banker]

This is very true and the people obsessed with the count debate often miss this concept

Yea. I know 3 of those.

[Three]

BTW, how do the results compare to EV?

This is an interesting question. I haven't simmed wonging but the casino conditions usually have me almost playing all most of the time (of course sometimes you wong out but you have nowhere to go so unless you are losing badly playing through it is usually the play as techniques to leave and comeback must be used in reasonable moderation. Fortunately much of the time you win or tread water while at a disadvantage. I don't play through when losing start to really add up). At my favorite times to play not many tables are open and at other times they try to keep them crowded. I only have a couple of predictable windows where wonging can be done as one would like to wong. These predictable windows may be short or for a couple hours. For this reason I use a slightly adjusted depending on the wonging done playall sim and I tend to come in way above that. A general overall approximation of wonging availability would not reflect what is seen on a particular day. I guess you were more referring to overall EV. I never tried to guess at overall average availability to wong for a sim to use for the reasons stated but my sense is I am coming in at EV if such a sim could be done accurately (That is assumptions on wonging availability and other interactions in the casino that would affect the way you wonging. Like if I have a nice win at a table I might wong out quicker than if I haven't shown them anything yet. I don't want to show them too much or win too heavily at any table or in any pit on a given day. That influences my wonging style as well. A general sim would almost certainly get these effects on wonging wrong).

[Three]

Interesting T3: Dont you ever tie?

Anyway going by this method my entries last year went like this:
116 winning days/casino and 62 losing days/casino for 178 casino-day entries. About 65% winning casino-day entries.

If you throw out trivial entries both winning and losing:
115 winners and 36 losers for 76% winning casino-days

As you can see the times I "tie" are removed from one to the next above. I call them trivial entries. I set the limit at less than $300 in the above context. Logging a trivial loss has value so my results reflect a heavy bias toward that. 1 trivial win and 26 trivial loses (ties as you put it). Most were a lot less than $300.

[Three]

T3: I've picked up some advice from you along the way on handling ties. Thus, my ties likely lean heavily on the side of total profit. Sometimes it's those little things that add up to a great deal.

Wins or losses of less than $300 are trivial. I like to book a loss if that is the result I am walking away with. You keep your session win rate low but your profits high. At a glance your file looks like you are a loser. For the same reason I like to keep my wins from getting too large. The delicate balancing act that grants longevity while being very profitable. I checked one of the places I play most for win for last year. While quite a large number it was probably a third of what I actually made there. Perhaps I am trying harder than I have to at preserving longevity.