How good is revere point count?

[Automatic Monkey]

Okay, I found the table I was searching for on another site. After reviewing it, my memory served me well and the table in the post indicates my initial position - Hi-Lo crushes Zen/Mentor in the game described....

Then there's probably a problem with the way you are using the data. Nothing "crushes" anything at this level. You're looking for a 5-10% difference at best.

 

[LovinItAll]

Then there's probably a problem with the way you are using the data. Nothing "crushes" anything at this level. You're looking for a 5-10% difference at best.

That's what I said in my original response, a >5% difference. "Crushing" is subjective, but all things being equal, 5-10% with little or no additional work is crushing in a BJ system, IMO (not comparing an L1 to an L3, of course). The difference between $567 and $656 an hour (or its relative amount) is significant to some people. I respect your opinion if it differs from mine.

Re: The way I'm using CVData - The only info I don't have is the ramp. That could be it, but I don't think it is. It isn't just the data, it's the author's comments as well, to wit (I paraphrase): "Hi-Lo crushes systems x,y, & z in these games".

Thanks, and I'll post the data in this thread when I'm done with my sims. I think users such as the OP will find it very helpful, though more experienced players may find it has little/no value.

Best ~ L.I.A.

 

[boneuphtoner]

Okay, I found the table I was searching for on another site. After reviewing it, my memory served me well and the table in the post indicates my initial position - Hi-Lo crushes Zen/Mentor in the game described.

Please show us that data.

Check out QFIT's Modern Blackjack Volume 2. In this masterpiece, there are literally hundreds of pages of data and charts comparing FELT (rounded reduced RPC using count per whole deck), HiLo, and REKO. I don't recall any of the data in this very comprehensive study showing HiLo beating FELT, and I would like to see the data you have showing HiLo beating similar Zen/Mentor.

 

[LovinItAll]

Please show us that data.

Check out QFIT's Modern Blackjack Volume 2. In this masterpiece, there are literally hundreds of pages of data and charts comparing FELT (rounded reduced RPC using count per whole deck), HiLo, and REKO. I don't recall any of the data in this very comprehensive study showing HiLo beating FELT, and I would like to see the data you have showing HiLo beating similar Zen/Mentor.

Okay...

Count Backula wrote a great article for another site comparing various counting systems. While he didn't use CVCX, I just loaded the given parameters for Hi-Lo and Zen to compare results (I didn't see Mentor in CVCX). They are not identicle, but the differences are, imo, statistically insignificant. Here are his results for Hi-Lo v. Zen (CVCX results in italics):

"Strip game. Pen: 4.5/6, S17, DAS, RSA, LS, 1:12 spread, play all, spread 1x50 to 1x600 (one spot), bankroll = $50,000, heads-up, 300 hands/hr, full indices. (1b+sims)

HiLo: $448/hr, RoR 6.6%, SCORE 40.34, N0 24,789
($444.39, 6.4%, 40.53, 24,673)

Zen: $415/hr, RoR 8.0%, SCORE 34.80, N0 28,737"
($417.65, 7.5%, 35.85, 27,896)"

From CB's data, I consider ~8% difference in win rate and an RoR improvement of ~15% to be absolutely 'crushing'. To me, not 'crushing' would be comparing Hi-Lo to FELT (CVCX, same parameters):
($447.75, 5.4%, 43.47, 23,005)

FELT is clearly better than Hi-Lo in this game, but if for some reason a player was not particularly risk-sensitive, learning FELT would yield no appreciable benefit. A more risk averse player might have a different view.

(No disrespect intended toward the FELT system. It may perform better in a different game - I've only read about it briefly on the QFIT website).

Best ~ L.I.A.

P.S. Many thanks to CB for permission to reprint his results.

 

[boneuphtoner]

Which version of Zen was used for this comparison? I ask this as the true edge version of Zen is known to be significantly weaker than the original version. According to the canned CVCX sims, true edge Zen is indeed weaker than HiLo. Traditional Zen, using count-per-deck, is much stronger, and easily beats HiLo.

 

[LovinItAll]

Which version of Zen was used for this comparison? I ask this as the true edge version of Zen is known to be significantly weaker than the original version. According to the canned CVCX sims, true edge Zen is indeed weaker than HiLo. Traditional Zen, using count-per-deck, is much stronger, and easily beats HiLo.

If you ran the canned sim in CVCX, I guess it's obvious that I used that, too. I don't know what CB used for his sim.

If you're saying that the versions of Zen that both QFIT and whatever sim CB used are the core of the problem, you may be right - I don't know. Maybe the original version of Zen can make up that much ground. I thought it was common knowledge that Zen didn't excel in shoe games, hence my surprise when I was taken to task for my original comments. Every other piece of data in CB's article checks out, and you would have to ask Norm why he chose to use what you're describing as inferior parameters for his application data.

If you'd like to run a sim in QFIT's app using Zen and the count-per-deck, I'm sure those interested in comparing systems would appreciate the update posted here. I would like to see the exact setup, too. Really, the statement "Traditional Zen, using count-per-deck, is much stronger, and easily beats HiLo" demands it, I think.

Best ~ L.I.A.

 

[zengrifter]

Okay...

Count Backula wrote a great article for another site comparing various counting systems. While he didn't use CVCX, I just loaded the given parameters for Hi-Lo and Zen to compare results (I didn't see Mentor in CVCX). They are not identicle, but the differences are, imo, statistically insignificant. Here are his results for Hi-Lo v. Zen (CVCX results in italics):

"Strip game. Pen: 4.5/6, S17, DAS, RSA, LS, 1:12 spread, play all, spread 1x50 to 1x600 (one spot), bankroll = $50,000, heads-up, 300 hands/hr, full indices. (1b+sims)

HiLo: $448/hr, RoR 6.6%, SCORE 40.34, N0 24,789
($444.39, 6.4%, 40.53, 24,673)

Zen: $415/hr, RoR 8.0%, SCORE 34.80, N0 28,737"
($417.65, 7.5%, 35.85, 27,896)"

I'd like QFIT's interpretation of these results. zg

 

[LovinItAll]

I'd like QFIT's interpretation of these results. zg

There may be some explanation why the data isn't conforming to the way others want to see it, but it should be clear that I haven't manipulated it. I'd like to make these points:

- In response to my post that said, "I thought I remembered that even Hi-Lo crushes Zen/Mentor in a shoe game.", you said, "You are wrong".

- I took the time to go back and find the data that supported my statement. It was posted on the site of arguably one of the best BJ writers and 'system developers', and if there's a procedure that the entire BJ community frequently uses that's named after someone, that 'someone' is, I think, regarded as reputable by default. Further, it was acknowledged by him as an excellent post, and there was no repudiation re: any of the results by said writer.

- I then ran the sim using the software that is touted as being excellent by the users here, including me.

On a site where math rules and where comments not supported by math are quickly shot down, there are no 'interpretations of the results'. There may be another simulation that can be run that will yield different data, but the results are the results.

I'm not trying to pick a fight, but if someone doesn't like the sims used, why don't they go back, input the data, and run a different one instead of waiting for someone else to chime in? If the original Zen system released 20 years ago was revised due to the level of difficulty implementing it, new players trying to decide on a system are probably going to evaluate the newer version.

The article referenced addresses the "what system should I use?" question. As such, players who are using a system that may perform better for them are encouraged to keep using it. If it ain't broke......

Here are some additional comments from CB's lengthy article:

"What system should I use?"

The answer for 90% of card counters should be HI-LO. There are only a few specific reasons not to use this count. The inability to divide properly in order to convert to the true count is one of them. The desire to get more of an advantage by adding the 7 as a +1, thereby switching to the unbalanced KO, is another (shoe games only, no tracking).

Here are the only reasons I can think of NOT to use hi-lo:

1. Inability to convert to a true count. Use K.O., or alternatively, use Red Seven.

NOTE: As strange as it seems, to increase your win rate may not be a valid reason. Hi-lo won't make you much less money than the most complicated count. As many others have stated (and I won't elaborate here), increasing your bet spread a couple of notches outweighs the use of any special counting system and/or side-count. However, .... a more complicated system does make more money, so it's up to the reader to determine whether or not the increased errors or fatigue is worth it.

2. To reduce Risk of Ruin, especially because of the use of a non-replenishable limited bankroll (less than 75k is a limited bankroll). If you only have so much money, and you're not betting big enough to be noticed, and importantly, you're skilled enough to use it, then your act isn't as necessary and you could concentrate on the game more. A more complicated system, when used properly, will reduce your RoR (while increasing win rate (WR) a little), and so therefore go ahead and "go krazy" with your crazy count.

3. You are smart, quick, a perfectionist and/or bored. Using a more complicated count can stimulate your brain if it needs stimulating. Some say it's good for you. Others call it mental masturbation and a waste of time that will tire you out too soon. In any case, don't think that it will increase your WR by 30% or anything silly like that. It won't. Increasing your bet spread (and wonging) will, as well as advanced AP techniques. Keep in mind that arguably the biggest blackjack winner of all time used hi-lo, and without full indices. It is also psychological to know that you're using a kick-ass multi-level system (sometimes with an ace side-count). If it makes you really tired, switch back to hi-lo. For example, several halves counters switch back to hi-lo when tired. Or drunk.

Generally accepted summary

Use HiLo. If you can do it easily, use AOII or HOII with aces side-count for pitch games, and RPC or halves for shoe games. If you play lots of pitch and don’t want to side-count aces, use HiLo for a simple system and Zen for a level-2. For a combination of game types, use HiLo. For a more complicated system (no ace side count), use AOII or HOII.

For unbalanced systems, use KO Full. For a level 2 unbalanced, use UBZII (SD and DD only).

This is only my opinion and I admit others have valid arguments against these choices, but I think APs should exclusively use only following counting systems:

1. Best overall simple (no side counts): Hi-Lo
2. Best level 2/2+ for shoes: Halves or RPC
3. Best level 2/2+ all-around: AOII or HOII or Halves or RPC
4. Best pitch count with aces: AOII / Hi-Opt II
5. Best level one count plus aces, for pitch: Hi-Opt I + aces (and the almost equivalent omega I and revere +/- plus aces)
6. Best simple “count every card” count: Silver Fox (for those who do better when counting most of the cards)
7. Best Unbalanced simple: KO
8. Best Unbalanced level-2: UBZII (pitch only)
9. Best pitch level-2 count (no ace side count): Zen

I recommend using the following counting systems (if you are not side counting Aces):

1. Hi-Lo – simple
2. RPC – level 2, good all-around but better at shoes
3. Halves – level 2+, better at shoes, but so good that it can still be used for pitch
4. Zen – best balanced pitch system without a side count, period
5. UBZII – best unbalanced count for pitch, period
6. KO Full – best all-around simple unbalanced system"

There's much more, and the article goes on to include dozens of sims comparing virtually all systems (except FELT. The article was written in 5/09). Also, systems like A-5 and KISS were not included.

To me, the article was the best/easiest-to-read comparison of counting systems I've ever seen posted online. YMMV.....

Best ~ L.I.A.

 

[bjcount]

Lovinitall

I had a conversation with CB some time last year about a similar scenario. I trust his sims to be accurate but I come up with different results using a different set of indices for Zen and RPC. The Hilo indices are as provided with CVData.
Although it's not questioned that limiting your use of indices to just the I18 and Fab 4 in 6d games is recommended, I learned the full set, so I use a full set of approx 94 indices.
The Zen and RPC indices used in the sims attached below are a full set of risk averse (RA) indices generated using CVData. I use RPC and a friend uses Zen.
All the bet spreads were optimized prior to running the sims for that respective strategy, no "standard" bet spread was used.

If your going to just count and grind it out then I see no reason that you shouldn't go for as many indices as you can if you play all. You'll look like an idiot hitting 13 v 4, but you'll know why your doing it.

I used the rules you quoted in your post above, "Strip game. Pen: 4.5/6, S17, DAS, RSA, LS, 1:12 spread, play all, spread 1x50 to 1x600 (one spot), 250 hands/hr, full indices. (1b+sims)".
My version will not do 300 hands/ hr, 250 is max.
The order below is Hilo, RPC (#1 BJC real play), and Zen

BJC

 

[tthree]

Wow!
Comparing win rates:
Zen/HILO = 376.57/303.06 = 1.24255923

That's a 24.26% increase in win rate! If 7.08% is absolutely crushing what is over 24%.

I notice your Zen indices are risk averse, that could be a big chunk of the difference. That should be less than 10% of the increase though. Maybe less than 5%.

 

[LovinItAll]

Hey BJC,

Thanks for the reply. As I think I mentioned, I brought CB's post up not so much as a point that I want to go to my grave debating, but more as a general comment re: CB's overall article and its usefullness in helping someone new decide on the direction they might wish to travel if they are trying to decide on a system. None of the info has made me inclined to change the systems I use.

That's a 24.26% increase in win rate! If 7.08% is absolutely crushing what is over 24%.

Slaughtering

On to BJC's comments::

Although it's not questioned that limiting your use of indices to just the I18 and Fab 4 in 6d games is recommended, I learned the full set, so I use a full set of approx 94 indices.


If your going to just count and grind it out then I see no reason that you shouldn't go for as many indices as you can if you play all. You'll look like an idiot hitting 13 v 4, but you'll know why your doing it.

Of course. If it doesn't wear someone out, imo they should be using whatever they can to maximize EV and/or minimize risk (or both, depending on their individual circumstance).

I used the rules you quoted in your post above, "Strip game. Pen: 4.5/6, S17, DAS, RSA, LS, 1:12 spread, play all, spread 1x50 to 1x600 (one spot), 250 hands/hr, full indices. (1b+sims)".
My version will not do 300 hands/ hr, 250 is max.
The order below is Hilo, RPC (#1 BJC real play), and Zen

BJC

Did you use CVData for the comparison? I think it only goes to 250 hnds./hr. Regardless, when I use CVCX, game as specified, 250/hr., I get a different result:

At $373.26, RoR of 6.6%, and a SCORE of 40.34 (not optimizing the betting ramp, btw), HiLo once again holds its own. I suspect that if I entered the data as you did for the other systems, they might very well shine.

I just can't stress this enough: I'm not trying to defend HiLo or dog another system. For me to even remember CB's post and then to actually be able to find it is more than I expected from my small brain.

'Crushing' aside, I think everyone agrees that whatever system a player can reliably use will offer the most EV for them. As CB pointed out, when comparing apples to apples, no system is going to offer a 20-30% advantage over another generally accepted 'good' system, even if it is an L1 (tossing aside A-5 and the like).

Take care,

L.I.A.

P.S. I'm using CVCX, as that's what Norm recommended.

 

[assume_R]

If your going to just count and grind it out then I see no reason that you shouldn't go for as many indices as you can if you play all. You'll look like an idiot hitting 13 v 4, but you'll know why your doing it.

I'm not sure I agree with this. Run 2 sims. First, include all indices. Then, just take out ones you aren't sure add much. For example, I'll use your 13v4 example.

After 4 million hours of play with a 13v4 index:
Win Rate = $46.89/hour, Std = $645.75/hour, SCORE = 52.72

After 4 million hours of play with always standing on 13v4:
Win Rate = $46.76/hour, Std = $645.59/hour, SCORE = 52.46.

So a player who always stood on 13v4 would be getting 99.5% of the SCORE of someone who knew the index, and only realized after 4 million hours (many lifetimes) of play. Before 4 million hands the winrates fluctuated too much to have a noticeable difference.

If the player just got a few more cards of pen, found a table with 1 less player, or played for a minute or 2 more a day, or wonged out of one more TC < -2 shoe (so that 13v4 index never came into play), that would be more worth it than learning all the extra indices combined.

To each his own though; I'm sure many would disagree with me on this :grin:.

 

[LovinItAll]

I'm not sure I agree with this. Run 2 sims. First, include all indices. Then, just take out ones you aren't sure add much. For example, I'll use your 13v4 example.

After 4 million hours of play with a 13v4 index:
Win Rate = $46.89/hour, Std = $645.75/hour, SCORE = 52.72

After 4 million hours of play with always standing on 13v4:
Win Rate = $46.76/hour, Std = $645.59/hour, SCORE = 52.46.

So a player who always stood on 13v4 would be getting 99.5% of the SCORE of someone who knew the index, and only realized after 4 million hours (many lifetimes) of play. Before 4 million hands the winrates fluctuated too much to have a noticeable difference.

If the player just got a few more cards of pen, found a table with 1 less player, or played for a minute or 2 more a day, or wonged out of one more TC < -2 shoe (so that 13v4 index never came into play), that would be more worth it than learning all the extra indices combined.

To each his own though; I'm sure many would disagree with me on this :grin:.

I agree with this in principle, but in some cases I think things begin to have either a cumulative or mental effect. For example, people who don't learn something because it's 'too much trouble' will often find that applying that reasoning to everything they do will start to erode their performance (EV or otherwise). A perfectionist might learn the correct strategy from jump street thinking they'll not use some plays situationally, but because of their predisposition, they can't overlook even tiny increases in performance (EV or otherwise) unless it's part of a different strategy (cover, et al).

In the above example, a perfectionist is likely to look for improvement everywhere (pen, fewer players, etc.) while the other person may be more likely to settle for less favorable conditions more frequently.

All things being equal, who would you rather have as an employee? I guess arguments can be made for both, but......

Best ~ L.I.A.

 

[assume_R]

I agree with this in principle, but in some cases I think things begin to have either a cumulative or mental effect. For example, people who don't learn something because it's 'too much trouble' will often find that applying that reasoning to everything they do will start to erode their performance (EV or otherwise). A perfectionist might learn the correct strategy from jump street thinking they'll not use some plays situationally, but because of their predisposition, they can't overlook even tiny increases in performance (EV or otherwise) unless it's part of a different strategy (cover, et al).

In the above example, a perfectionist is likely to look for improvement everywhere (pen, fewer players, etc.) while the other person may be more likely to settle for less favorable conditions more frequently.

All things being equal, who would you rather have as an employee? I guess arguments can be made for both, but......

Best ~ L.I.A.

Oh I very much agree with you on that. The attitude is indeed important. But for me, for example, I have spent my time learning multi-parameter counts and learning how to play several different casino games. At each game, though, I find the important indices, or plays, or information, and focus all my attention on them.

 

[QFIT]

Zen as originally published was a great strategy. When it was republished with "true-edge" logic, the way that most people would play given that strategy, would result in dramatically lower profit. I discussed this with Arnold at the time, and added support to my software to allow betting by half-counts as it would be required to bring the results back up. Unfortunately, I discussed true-edge with numerous players over the next several years, and frankly, they did not understand what Arnold was trying to do and did not adjust their play accordingly.

I still believe that the original Zen was a great strategy, although it was missing a few important indexes. I remain a larger fan of RPC as I believe it is simpler as Tens and Aces are counted the same. This is not an insignificant difference. IMO, RPC is vastly faster when counting two cards at a time. And, I believe this is very important as speed is how you get to the long run. This is a concept I aim at in my book many times. To get to the long-run, you must play a huge number of hands. In my mind, speed is important because it gets you to the long run more quickly, as well as the simple fact that more hands means more profit.

True-edge betting was a noble attempt at simplification. But, it was developed before Kelly-betting was fleshed out and requires Wonging for effective use.

 

[bjcount]

Hey BJC,

Did you use CVData for the comparison? I think it only goes to 250 hnds./hr. Take care,

L.I.A.

P.S. I'm using CVCX, as that's what Norm recommended.

Yes, I used CVData and that's correct, it only goes to 250 hands/hr..... oh how I wish I can get a few dealers to run out the cards that fast on a regular basis.... I've only found 1 in 3 years that ran out (7) 6d shoes in 1 hour heads up.... and that includes shuffle time... and now she's gone....

There's 1 or 2 dealers floating around here that can do 5 shoes in an hour heads up ..... and thats just pushing the 200-250 hands/hr. point.

Other than that, your lucky to find dealers running out 3-4 shoes/hr. Dam turtles....

BJC

 

[QFIT]

Fastest dealer I've ever encountered with multiple players was in Puerto Rico. The cards seemed to appear out of nowhere. Fastest dealing heads up was at the old DI the day after Hughes died when I was the only player in the entire casino. No idea how fast it was. We got into a rhythm I’ll never see again – and chatted the whole time.

Oddly, once you have practiced enough, it can be easier to count in such situations. There simply is no time to forget the count.

 

[LovinItAll]

Oddly, once you have practiced enough, it can be easier to count in such situations. There simply is no time to forget the count.

Plus, if the floor saunters over, you can give them the "damn she's fast!" look/comment, sort of shake your head a little startled like it's all you can do to add to 21 because of her speed and focus on the game. Spreads seem to go less noticed, too.

At the risk of sounding sexist....

Usually the female dealers who try to go too fast would be better off slowing down a notch. One in particular comes to mind, and the 'card down' call once/hour got a little old. The cards usually ended up about a foot apart, too. She was promoted (obv).

One the other hand, there's a woman at the same place who blazed in the D/D pitch game - she was just a machine. I was getting chilled one night and started intentionally slowing her down despite it not being in my best interest. The slaughter went unabated - maybe if I hadn't been so stupid it would have turned around.

On the third hand, the dealers who have that slooooow, high pitch make me want to pluck the cards out of the air and throw them back. These are usually the same folks who have excruciatingly long stories that require stopping the game periodically for dramatic effect.

 

[zengrifter]

1. Inability to convert to a true count. Use K.O., or alternatively, use Red Seven.

Rensey's KISS-3 is best-in-class. zg

 

[zengrifter]

If you can do it easily, use AOII or HOII with aces side-count for pitch games

No, except for that "mental kick" you were saying.
Unless you are using a secondary count overlay (2s5s +1 vs As -2), you
wind up working harder for the same result (at best) as ZEN or Mentor. zg