CVDATA Questions

[Duc939]

Hi all,

I've now been using CVDATA and CVCX for a couple of months and have loved it, after finally getting the time to read through the entire manual of CVDATA I have the following questions, it would be great if Norm or anyone can chime in. Apologies in advance for the lengthy post, and if there's any questions where you feel it's outside the scope of this forum to answer (such as reference to side counts perhpas) please feel free to point me in the right direction for research / reading. Out of below it's the first point which really has me second guessing, I can't seem to get results which make any sense to me. Here goes-

1. I can’t seem to figure out how to correctly use the ‘Sessions’ tab and function in CVDATA. For example, using CVCX I used a Playing Strategy and derived a Betting Ramp using the ‘Custom Bets’ column with a 5K bankroll which gave me an 8.7 Risk of Ruin. Using CVDATA I started a Multi-Tracking sim, I used the same Playing Strategy, all the same playing rules, and created a Betting Strategy which perfectly mimicked what I had derived on CVCX, Wonging options were also set identical between CVCX and CVDATA. I ensured Unit Size under ‘Betting’ was correct ($5) and went to ‘Sessions’ under ‘Nuances’, here I typed ‘1000’ into starting bankroll (1000 units = $5,000), I also set session target to 1000, I selected ‘No maximum session length’ and ‘minimum session length’ set to ‘1’. I selected ‘shuffle after bankruptcy’. So if I understand this correctly I am asking the sim to tell me how many times I would profit 5K from a 5K starting bankroll using my betting and playing strategy, now CVCX had me at a ROR of 8.7% so I would be expecting to see that I lose my entire bankroll roughly 8% of the time given my goal is to double my bankroll, however after running the sim I received the following values. BANKROLL=5,000 TARGET=5,000 MAX ROUNDS=100 (even though set to ignore) SESSIONS=463,147 WON=99.85% LOST=0.15% PUSHED=0 HIT TRAGET=99.85% BANKRUPT-0.15%. I understand the 8.7% ROR isn’t based off doubling bankroll or going broke but only going broke 0.15% of times before doubling a 5K bankroll with an average bet size of $20 and a ROR of 8.7% seems too low? When I changed my TARGET to 10,000 units I totalled WON 30% and LOST 70% under the same conditions. The sessions data also seems to sometimes only record some seats and not others.
2. I’ve noticed when running a CVDATA and CVCX sim with all the same parameters including betting strategy and playing strategy and rules, that some values come out slightly different, for example Win Rate and ROR. Obviously no two sims will get the exact same value, but do you have advice on which one would generally be more accurate? I would assume CVDATA as it has been run with a particular betting strategy whereas CVCX the betting ramp was adjusted AFTER the sim was run?
3. My next question relates to ‘Multi-Parameter’ side count on the Index Generation, I don’t fully understand this. For starters there seems no way to select a ‘T or X’ only ‘A-9’ in the scroller. Now if I set ‘1’ in the scroller which I assume is ‘A’ and select to run it for HDD it gives me indexes of (0 for a 7 against a dealer 5) and (2 for an 11 against a dealer 5), are you able to give any guidance on how this is actually used?
4. When selecting “last round bet’ under Advanced Betting Strategies option, I’m assuming you are selecting the number of cards prior to the point of where you have selected the cut card to be placed under Rules?
5. When editing a Betting Strategy under ‘Last Hand” there are options for ‘count<pivot’ and ‘count>pivot’. What does this mean? I noticed the ‘pivot’ come up in a floating advantage table too.
6. On page 65 of the Reference Manual when describing Hand Depth Won/Lost you say – “When using an effective strategy, the amount bet and won should increase as the depth increases”, however the value under ‘bets’ and ‘won‘ dramatically decrease on my sim ran, this is using a hi-low system which gave me a SCORE on CVCX of 39.40
7. On the Dealer Totals chart in CVDATA sometimes when I run sims I get values of amount of times the dealer finished on hands lower than 17, it only happens sometimes and the only thing I changed between the sims was how many payers sat at the table. All my sims are run with no holecard European rules and no bonuses. For example it says the dealers upcard is a 2, and he finished on a total of 5 four times. How can he finish under 17 with no holecard and no bonuses forcing to draw a card after busts? The same table is showing that the dealer busted on a 10 double the amount of times he did on a 6, this just doesn’t seem right?
8. In the Negative Bankroll Excursions table I’m unclear what the rows are reflecting. >50 hands? >50 sessions? Etc. For example under <50 it says 2,134,454, this can’t be saying the negative swing resulted in 2 million units under the previous high within 50 hands? In the reference manual the description doesn’t seem to match the table I’m getting.
9. Can you advise me what is meant by “block size” and “block numbers” when viewing the Results Log?
10. I don’t see ‘Bankroll Log’ under Tools even though I have Session Data and ‘keep a log’ selected. Only Results Log and Data Log
11. I ran a Multi-Sim with two players, player one used a custom betting strategy and player two used the Flatt Betting strategy included. When viewing Hands Log it shows seat ONE and seat THREE, why not ONE and TWO? And the TC for seat three is consistently in the 20’s or 30’s. The TC for seat ONE seems accurate though.
12. When I click on reports after running a sim I can select various reports however under “Players Reported” is blanked out and I can’t select any seats and therefore generate or print the reports.

Thanks for reading this far, looking forward to having a better understanding and getting more out of this software.


Duc.

[Norm]

1. The target is not amount won but amount you want to end up with including the bankroll. If they are the same, the session ends immediately if you win the first hand. You can look at the session data and see that the average session length is tiny.

2. CVData is accurate. CVCX will give the same results with the same betting strategy, unless you use the after sim wonging and multiple hand functions which use estimation techniques. .Using the CVCX pre-sim wonging function is accurate.

3. Very few people use MP counts. Only Humble’s Hi-Opt book comes to mind where there is any discussion. Haven’t heard of anyone side counting tens for an MP count. MP tables are normally used to improve playing efficiency by including the effect of the cards counted as zero. For example, in Hi-Opt II, 8, 9 and Ace are counted as zero. The following occurs:

• Calculate the number of excess eights (may be negative) in the remaining cards
• Multiply this number by the index value in the multi-parameter table for the current hand.
• Temporarily subtract this from the running count
• Recalculate the true count according to the current strategy's instructions for playing purposes only.

Example: Suppose you have 9 vs. 2. The deck has two excess eights. The Hi-Opt II MP index for 9v2 with a side count of eights is 2. Multiply that 2 by the 2 excess eights. You now temporarily reduce your RC by 4 making it less likely that you will double. This makes sense since the excess eights make it easier for the dealer to get a 20. This would have been missed by the normal Hi-Opt II count since it ignores eight.

4. Yes. The spinner allows you to set the maximum number of remaining cards that indicate a last round is coming.

5. Pivot is a term used for unbalanced counts, like KO. It is always zero for balanced counts.
6. Not sure I understand this one. CVCX doesn’t provide these stats. They are in CVData results.
7. If the players all bust or get blackjacks, the dealer doesn’t draw.

8. I’m rewriting this. It’s very confusing.

9. These charts are divided into blocks of rounds. For example, if block size is 500, each data point in the chart shows the result of 500 rounds. For a block size of 1, each data point is a single round. You can then change the block number to step further into results.

10. If you mean the Hand Log, it is only generated when Massive Data is selected on Nuances-Data and is not generated using the standard simulator. The Mult- and shuffle simulators generate it.

11. I’d have to see this one. I assume the first player plays two hands requiring two seats. The Flat Betting strategy is set to use a running count instead of a true count. Create a new betting strategy with running count turned off.

12. This should only occur if there is only one player or with the standard simulator where data for only one player is collected.

[Duc939]

Hi Norm, thanks for the reply.

Regarding the ‘Sessions’ I’ve been wrecking my mind, what you said cleared some things up but it still hasn’t been making sense at times, after running a bunch of tests I believe it has randomly being giving results for incorrect players which is why the figures just didn’t make sense at all, such as huge bankrupt rates for players with 8% ROR, even after adjusting the targets. I’ve detailed some of the sims below.
All sims were conducted with a combination of players consisting of one of two strategies –

Player A – Uses a modified playing strategy with indices and a betting strategy consisting of two hands above a TC+3, ROR is 8% and wonging is used.
Player B – Uses Basic strategy correct for game rules and a flat betting system, should be playing at about a 0.5% disadvantage.

All sims are run under multi, with a 1,000 unit starting bankroll (5K). No maximum session length and a minimum session length of 1 round.

For sim one I elect to have FOUR players, all are Player B configuration. Session target is 1100 units After running the sim the sessions data shows FOUR seats, all with about 40% of hitting target and 60% going bankrupt, so far so good. If I click on the ROR calculator I am able to load data from all FOUR players, although sometimes it only shows three (ie. Player 2 is greyed out) until I close and re-open the calculator, a small bug.

For sim two I have TWO players and a target of 1100 units, the first seat is Player A configuration and the second seat is Player B. This time after running the sim I see only one seat under sessions data not two. It is clearly showing Player A, he has a 95% win rate. The ROR calculator will only load from one seat and it is Player A with an 8% ROR. Where’s details for the second player?

The third sim seems buggy I think and this is what I was getting before every time which was confusing me. Same set up as above. Player A configuration in seat 1 and Player B configuration in seat 2. I ran this sim with a 2000 unit target (so 10K total from 5K). After running the sim I again see only SEAT 1 under sessions data and no second seat, so I assumed this must be player A configuration as he was in seat 1, it however shows a 100 percent bankruptcy rate. I believe it must be showing Player B who was actually in seat 2 flat betting and not counting or wonging. When I click on the ROR calculator it will only load Player 1, which seems to be Player A configuration with a low ROR.

The fourth sim for testing I run FOUR players. Seat one and two as Player A configuration and seat three and four as Player B configuration. Target is set to 1050 units from 1000 starting. After running the sim only TWO seats are displayed under the sessions data, not FOUR. Seat one bankrupts 1% of the time and is clearly Player A configuration, seat 2 bankrupts 47% of time and is clearly Player B configuration who was in seat three and four when I set the sim up. When I select the ROR calculator Player 2 is greyed out, Player 1 loads from Player A configuration and Players 3 and 4 load from Player B configuration. Seems odd and a little random.

All of the above sims I have saved and archived. If I’m not mistaken it seems when setting up players which play multiple hands it is perhaps messing with the sessions data a bit.

Cheers.

[Dog Hand]

Hi all,

<snip>[*]On page 65 of the Reference Manual when describing Hand Depth Won/Lost you say – “When using an effective strategy, the amount bet and won should increase as the depth increases”, however the value under ‘bets’ and ‘won‘ dramatically decrease on my sim ran, this is using a hi-low system which gave me a SCORE on CVCX of 39.40
<snip>

Duc.

Duc,

This looks like an example of the "Cut Card Effect" or CCE, in which the RC, TC, and EV all plunge as the "penultimate penetration", that is, the penetration on the second-to-last round, approaches the fixed-depth cut card position.

Here's a simple example to demonstrate the CCE. Say you're playing B.S. on one spot heads up against the dealer on a single deck game (with an IBA of -0.45%) where the dealer fixes the penetration by inserting a cut card after the shuffle and cut.

If she places the CC right after the fifth card from the top (for a whopping 9.6% pen), then after she burns the top card only four cards will remain before the CC. Since the first round is guaranteed to use at least four cards, she will shuffle after every round. In this case, the EV on every round will be the same -0.45%, so here there is no CCE. By the way, you'll also see the same -0.45% if the CC is instead inserted after the second, third, or fourth card.

Now let's say the dealer likes you, so she gives deeper pen: she inserts the CC after the sixth card, giving a magnificent 11.5% pen... hey, that's nearly 2% deeper pen, so you should be happy, right?

Now, if the first round of the shoe uses exactly four cards, then after that round the CC will still not have come out nor will it be the top card of the pack, so she'll deal a second round. On the other hand, if the first round uses five or more cards, she'll deal only that first round out of the deck, as before.

Clearly, on Round 1 of every deck, your IBA will be the -0.45% as before.

What will be your IBA on the occasional Round 2? To answer this, we need to consider why the first round used only four cards. This will happen if

1. a BJ occurs for either you or the dealer (or both), or

2. both you and the dealer have pat 2-card hands.

In either of these two cases, the RC at the start of Round 2 will never be positive (using HiLo), and will only occasionally be zero: for case 1, for the RC to be zero, the non-BJ hand has to be two smalls, otherwise the RC will be negative. For case 2, the only time the RC will be zero is if you both are dealt 8-9, any other situation will give a negative RC.

So, on the occasional Round 2, rarely the RC and TC will be zero, so your EV will be approximately -0.45%. Most of the time, the RC and TC will be negative, so your EV will be lower than -0.45%. Therefore, on average, on Round 2 your RC and TC will be negative, and your EV will be worse than the off-the-top -0.45%. Since your first round EV is still -0.45%, you can see that your overall EV for this game will be lower than -0.45%.

If you sim this game and look at the results by hand depth, for a depth of 1 (thus, Round 1 after the dealer burns one card) you'll see the RC and TC are zero and the EV is -0.45%. For a depth of 5 (on the occasional Round 2), you'll see the RC and TC are both negative, and the EV is much less than -0.45%. Thus, the CCE causes these values to plunge as the depth approaches the CC depth of 6.

Thus, paradoxically, we see that increasing the penetration actually lowers the player's EV: that is the CCE.

Granted, this example is somewhat contrived, but the CCE is present in any game where the dealer fixes the pen with a cut card, though its effect on the overall EV is lessened as the number of decks increases.

Rather than using a CC, if the dealer deals a fixed number of rounds, say six rounds for a heads up single deck game, then the B.S. EV will be exactly the same on every round, so no CCE is seen.

Hope this helps!

Dog Hand

[Dog Hand]

Oops!

In paragraph 1 of my essay above, change "the penetration on the second-to-last round" to "the penetration after the second-to-last round".

Dog Hand

[Norm]

If you see something that doesn't look right, in the configuration, go to Tools-Export, create an export file and email it to [email protected].

[Norm]

Ahh, good catch. I didn't even have to change the simulator. I just deleted a line of code in the user interface. Install www.qfit.com/update5165.msi. You do need to sim more rounds.

[Duc939]

Thank you for that explanation it was quite well explained.