Page 1 of 2 12 LastLast
Results 1 to 13 of 22

Thread: Wolverine: SCORE vs Win Rate and s.d

  1. #1
    Wolverine
    Guest

    Wolverine: SCORE vs Win Rate and s.d

    Don,
    I've got CVCX running on the preferred BJ game I play with indices (I18+, of course!) and have the bugs worked out about calculating TCs, etc.... I started playing with bet schedules and ran into some data that I didn't understand fully. Please help.

    First, the DD game we're discussing is what it is. This chain of casinos isn't going to offer a different DD game, or change the penetration. I totally realize that "shopping" for better penetration would help bring the SCOREs up. But let's leave that suggestion out of it for now. Thank you.

    Let's assume a $100 unit. If a bet schedule is laid out as follows: TC<2 is 1 unit, TC=2 is 2 units, TC=3 is 3 units, TC=4 is 4 units and a TC of 5+ gets 6 units. The win rate for this game is $129.00 per hour, the SCORE is about 36.00 and the s.d. per hour is 21.62.

    If we tweak the betting schedule to be the same spread, but more condensed, some funny things happen. The new schedule would be: all negative TC get a 1 unit bet, TCs of 0 or 1 get a 2 unit bet, and then we hit the TC of 2 with a 4 unit and any TC of 3+ with a 6 unit bet. Obviously, we'd bet 2 units off the top since the TC is 0. Now, the win rate per hour goes up to $197 per hour, the SCORE is not much better at 38.00 and the s.d. per hour has grown dramatically to 32.05.

    Here is where I need your expertise. Since the SCOREs are nearly identical (a 5% difference) but both the s.d. and win rate went up by roughly 50% in the second bet schedule. It appears that although more money is hitting the table when I have an advantage, it is making the game an extreme roller coaster ride! Is the small increase in SCORE (5%) worth the greater risks taken with this more aggressive betting ramp?

    And what if we are willing to really get aggressive and make a betting ramp like this: TC <=1 bet 1 unit, TC of 2 gets 2 units, TC of 3 gets 6 units, and TC >=4 is two hands of 5 units each. I realize that the spread changes (effectively a 1-10 rather than a 1-6) so the comparison is only valid in determining if the schedule is better (which it should be due to the larger spread). The SCORE of this ramp is 44.80 with a win rate of $205.50 per hour and a s.d. of 30.72. If seems this ramp helps bump the SCORE and the win rate, while dropping the s.d. That would seem to be a good thing, right?

    I'm just trying to understand how win rates, s.d. and SCORE all work together. Plus, determine the best way to attack a given game. I've read that some players (pros, I guess) won't play a game unless they can get a SCORE of 50+. Is that a reasonable expectation for a DD play all approach?

    Thanks. I'll await your reply and hopefully I'll understand how to use this powerful SIM software to help fine tune my game.

  2. #2
    Don Schlesinger
    Guest

    Don Schlesinger: Re: SCORE vs Win Rate and s.d

    > Let's assume a $100 unit. If a bet schedule is laid
    > out as follows: TC

    Something didn't print here.

    > If we tweak the betting schedule
    > to be the same spread, but more condensed, some funny
    > things happen. The new schedule would be: all negative
    > TC get a 1 unit bet, TCs of 0 or 1 get a 2 unit bet,
    > and then we hit the TC of 2 with a 4 unit and any TC
    > of 3+ with a 6 unit bet. Obviously, we'd bet 2 units
    > off the top since the TC is 0. Now, the win rate per
    > hour goes up to $197 per hour, the SCORE is not much
    > better at 38.00 and the s.d. per hour has grown
    > dramatically to 32.05.

    > Here is where I need your expertise. Since the SCOREs
    > are nearly identical (a 5% difference) but both the
    > s.d. and win rate went up by roughly 50% in the second
    > bet schedule. It appears that although more money is
    > hitting the table when I have an advantage, it is
    > making the game an extreme roller coaster ride! Is the
    > small increase in SCORE (5%) worth the greater risks
    > taken with this more aggressive betting ramp?

    You're taking the SCORE name in vain! :-) SCORE assumes that you bet optimally. It has ONE ROR (13.5%), and all results are "normalized"; that is, higher is better ... period.

    It sounds like what you're doing is simply playing around with different sub-optimal ramps and spreads, which lead to different RORs. I can't tell you if taking more risk is worth it, to increase hourly win rate, but you might simply examine how much extra s.d. or ROR you get for a certain extra dollar amount and ask yourself if the tradeoff is "reasonable" to you.

    > And what if we are willing to really get aggressive
    > and make a betting ramp like this: TC =4 is two hands
    > of 5 units each. I realize that the spread changes
    > (effectively a 1-10 rather than a 1-6) so the
    > comparison is only valid in determining if the
    > schedule is better (which it should be due to the
    > larger spread). The SCORE of this ramp is 44.80 with a
    > win rate of $205.50 per hour and a s.d. of 30.72. If
    > seems this ramp helps bump the SCORE and the win rate,
    > while dropping the s.d. That would seem to be a good
    > thing, right?

    Again, you're speaking out of turn. You can't increase the win rate and drop the s.d. and still be playing the same game optimally.

    > I'm just trying to understand how win rates, s.d. and
    > SCORE all work together. Plus, determine the best way
    > to attack a given game. I've read that some players
    > (pros, I guess) won't play a game unless they can get
    > a SCORE of 50+. Is that a reasonable expectation for a
    > DD play all approach?

    Depends on rules, pen, etc. A $50 SCORE isn't as common as it used to be. But, be clear on this: You can't "make up" a bet schedule and then call the win rate that it produces a "SCORE." It's a win rate! (Reread "More on SCORE," BJA3, pp. 177-183).

    > Thanks. I'll await your reply and hopefully I'll
    > understand how to use this powerful SIM software to
    > help fine tune my game.

    Do this: Tell me your bank, or tell me your desired unit ($100?). Define the game precisely, leaving nothing out. I'll play around with things a little.

    Don

  3. #3
    Wolverine
    Guest

    Wolverine: Thanks, but...

    Don,
    Thanks for the offer to do some figuring, but I'm not comfortable sharing that specific information on a public board. I've been playing some more, but the learning curve with CVCX is steep!

    I understand exactly what you are saying about the use of the term SCORE. I apologize for using it to evaluate different bet ramps when it wasn't meant to be used that way.

    Sooooo....if I can't use the SCORE spit out by the CV suite, then what do I use? Just the win rate and the s.d. and determine if the RoR is too big with my bankroll?

    I've run a custom sim on CVCX to help evaluate an optimal ramp for the 1-6 spread, play all I wish to employ.

  4. #4
    Don Schlesinger
    Guest

    Don Schlesinger: Re: Thanks, but...

    > Sooooo....if I can't use the SCORE spit out by the CV
    > suite, then what do I use? Just the win rate and the
    > s.d. and determine if the RoR is too big with my
    > bankroll?

    Yes. And, if memory serves me, if it isn't a genuine SCORE, Norm writes "C-SCORE," for "Custom SCORE," which means that you have changed some of the basic premises that the original SCORE is based upon.

    > I've run a custom sim on CVCX to help evaluate an
    > optimal ramp for the 1-6 spread, play all I wish to
    > employ.

    That's fine.

    Don

  5. #5
    Wolverine
    Guest

    Wolverine: Yes, but

    On the Configuration and Results print out from CVCX, the Performance DI and c-SCORE are given when looking at optimal and custom bets. However, upon running (multiple) CVData sims, the Summary page lists plainly the DI and SCORE for a given set of circumstances. This latter tool is the one I was tweaking the bet schedule on and getting the win rate per hour, s.d., and SCORE info quoted in my first post without changing the playing strategy. If it is something other than a true SCORE, then I certainly apologize, I was just quoting from the output. I figured that comparing SOMETHING by changing the bet schedule would help me to understand what different bet schedules would do to win rates, s.d., and SCOREs. However, I found that changing the bet schedule only served to confuse me as to WHAT to compare (win rate, s.d., or SCOREs). Hence my question.

    Any thoughts?


  6. #6
    Don Schlesinger
    Guest

    Don Schlesinger: Re: Yes, but

    > Any thoughts?

    I can only repeat what I've already told you: If you choose a bankroll and a spread, optimal betting will give you a unit size and a SCORE. If you choose a unit size, spread, and optimal betting, you can find what bank you need, and you will get a SCORE.

    You CANNOT choose the bank, spread, AND the unit size and still bet optimally. What you get as output isn't a SCORE; it's an hourly win rate. So, you decide if you're happy with that win rate for the ROR that you assume.

    Don

  7. #7
    Wolverine
    Guest

    Wolverine: Thanks

    Being new to the whole simulation, index plays, RoR, and optimal bets, etc... makes it pretty hard. I'm sorry if I seem obstinate (sp?) or stubborn, I'm trying very hard to understand these concepts and what they mean to my individual situation. Since I've never compared SCOREs, win rates, s.d., or any other measure of performance, I'm very unsure of how they relate to other games, situations, etc... For example, if I told you my car went from 0 to 60 in 14 seconds, unless you had some frame of reference, that number wouldn't mean anything. I understand you have created SCORE to help give a frame of reference, since the parameters are fixed. That helps decide between different systems and types of games. But when evaluating a SPECIFIC game, I guess the win rate and s.d. and associated RoR will take over evaluation of any bet schedule. Therefore, SCORE won't work to look at that. I understand the optimal bet schedule concept you outline in your previous posts, but it doesn't fit into the real world casino experience very well. That is why I am trying to juggle it slightly and optimize the real world results. You can't jump from 2 units at +1 to 5 units with +2 TC and then to 6 units on the next higher TC--and live to play tomorrow! Yet, that is what the optimal bet spread called for. I was trying to find a gradual ascent up the betting ramp as directed in BJA3 (parlaying the wins, holding firm on pushes, and not increasing after a loss---but concentrating on the parlaying side of that advice with the bet ramp). Hence, I set up a 1-2-3-4-6 and 1-2-4-6 schedule and was trying to compare which schedule was better. I was also trying to get a benchmark (hence the SCORE thoughts) on what this game is so I can plug in accurate win rate and s.d. to get accurate RoR and results data back.

    I'll keep working on it and ask more questions when I run across them. That is the beauty of this group and you Don, the answers are out there. I appreciate your input and thoughtful responses.

  8. #8
    Don Schlesinger
    Guest

    Don Schlesinger: Let's try it this way

    > I'll keep working on it and ask more questions when I
    > run across them. That is the beauty of this group and
    > you Don, the answers are out there. I appreciate your
    > input and thoughtful responses.

    You didn't respond to my invitation to tell us the precise game you're playing (don't leave out a thing).

    Then, tell us either the unit you want to use, the bank you have, or the ROR that is acceptable (you only get to choose two out of three!) and the spread that you think you can pull off. We'll work this out for you.

    Finally, understand the following: In finance, many people swear by the Sharpe Ratio, which is a risk-adjusted-return measure (see BJA3). When comparing your different scenarios, assume as a base what a flat-betting BS player might achieve, both as e.v. and s.d., and then see what your proposed approach does to both of those figures.

    As a ridiculous example, suppose you increase e.v. by $1 but double s.d. That would be a dumb move. Suppose, instead, that you double e.v. and increase s.d. by only, say, 20%. That would be a spectacular move. Get the idea? The goal is to maximize the Sharpe Ratio, because so doing will also maximize the SCORE, or the custom-SCORE.

    Don

  9. #9
    Wolverine
    Guest

    Wolverine: Decline

    Thanks again Don. I put in a previous post that I am uncomfortable putting down such specific information on a relatively public board. Not an issue with you or the "regulars" but not sure I trust this place enough yet since I have seen some interesting things in posts.

    That last explanation was what I was looking for: how to take a SINGLE game/condition and look into the bet schedule. To me, the bankroll is fixed and the spread can easily be 1-6 and I might be able to get 1-10 by using 2x5 at the top. However, 10 units is too much for my bankroll according to a 2% (of bankroll) rule (I believe that is Kelly, no?) I have read in the past. So, 1-8 is the absolute MAX bet spread I should be playing, and it would probably be in the form of 2x4 to limit risk. However, when I do the calculations in CVCX, I am not happy with the RoR associated with that 2% max criteria. So, I am willing to endure less growth of the bankroll in order to minimize the risk. That is why I am not interested in the EXACT optimal bet schedule that the software will provide. As I have said before, my goal is not necesarily to GROW my bankroll for profit. I play BJ for fun, enjoy the comps, and as a hobby. If I break even and had a great vacation, awesome. If I lose a little, so be it--it is the way the cards fall. I just prefer not to get killed, as it would have then been cheaper to BUY a vacation rather than LOSE it!

    What I will tell you is this: I use hi-lo, Ill-18 + a few more indices, play all, DD, S17, DAS, NS and my 400 unit bankroll will not allow me to spread beyond 1-8 and I prefer 1-6 from the risk standpoint as outlined above. My bankroll is replenishable, but FIXED for the duration of the trip. According to the bet schedule CVCX creates, I slightly overbet my negative and 0 bets (the unit it picked is just a bit below the table limits found on DD, S17 games).

    From what I have been able to learn by studying the sim software data: any + TC is good, but at TC+2 or higher, I have the edge. Therefore, I am trying to minimize my bets at +1 or lower (only 1 to 2 units--2 units as a camoflauge play occasionally and at TC+1), and getting as many units out on the table as I can (within the Kelly criterion and being "welcomed back" to play/win again) when the TC gets to +2 or above. In setting up the different ramps, I was just looking for confirmation as to how to EVALUATE the data I am generating. Like you pointed out, doubling the win rate and only increasing the s.d. by 20% would be a great benefit--DO IT! If you only increase the win rate by 50% and double the s.d., probably not a good change--back to the drawing board. So, if I am listening (and I am trying)--find a win rate, s.d. and c-SCORE for a flat bet 1 unit player at this game. That then becomes my benchmark to compare other bet schedules against.

    (I haven't run the sim yet, but I would assume that the flat bettor would have a negative win rate and a low s.d.)

    Thanks Don. I appreciate your help.

  10. #10
    Don Schlesinger
    Guest

    Don Schlesinger: Re: Decline

    Two quick comments: 1) Why do you say that you increase your bet at +2? You have a clear edge at TC = +1 for the game you mention, and you should certainly increase your bet there.

    2) With a given bankroll, why do you say that you can't spread 1-10? You can spread anything you can get away with, but the UNIT size will be somewhat smaller. It would appear that you are not only fixing the bank but also fixing the unit size, as you try different spreads, which is a mistake.

    Don

  11. #11
    Wolverine
    Guest

    Wolverine: Understood, thanks

    > Two quick comments: 1) Why do you say that you
    > increase your bet at +2? You have a clear edge at TC =
    > +1 for the game you mention, and you should certainly
    > increase your bet there.

    Don, you are right, at TC=1 there is an edge in my favor. About +0.60%. I've played with it and found that just going from 1 unit to 2 units here helps the EV and barely moves the s.d. up, so a good suggestion to go ahead and increase at TC=1. Thank you. I did not realize how much this small edge would help the win rate. I assumed (and that is why I'm testing and tweaking, to learn about the math) that the risk of increasing my bet with such a small edge was going to be a much larger increase in s.d. and RoR since it was so close to being a 50-50 split. Turns out to not be so. Thank you.

    > 2) With a given bankroll, why do you say that you
    > can't spread 1-10? You can spread anything you can get
    > away with, but the UNIT size will be somewhat smaller.
    > It would appear that you are not only fixing the bank
    > but also fixing the unit size, as you try different
    > spreads, which is a mistake.

    I agree, but my bank is fixed for the trip and the next sentence will tell you why the unit has to be fixed too. Okay, I'm going out on a big limb here: I hope I don't regret this later because I've NEVER gotten heat in LV [knock wood]. The big Strip casinos that offer this game all have $100 minimums. Therefore, my smallest unit has to be $100. By CVCX, my "optimal unit" is calculated as $87. Since I can't bet $87 at a $100 minimum table, that unit doesn't work. I HAVE to fix my lowest unit. And since I am taking on greater risk on the bottom side of the table (when I don't have the advantage), when I spread up to $800 on the top end to try to make up for a bet at Full Kelly (2%), the RoR goes up to over 20%. I am slightly uncomfortable with the Kelly (?) & BJA3 risk of 13.5% of losing my entire bankroll. When it gets around 20%, I am petrified. I'm trying to get it down below 10%, but cannot with any bet schedule. (I can get it to 11.3% if I empoy a 1 unit <=+1; 2 units at TC=2 and 3; 4 units at TC=4 and 5; and 6 units at TC>=6.) I Wong out when the TC goes negative after playing a certain amount of time to minimize the downside, but I have to expect a "play all" approach and try to Wong out when it works out. So, I max out my bet at $600, which keeps the RoR down given the increased minimum bets.

    As I said before, I'm not a pro. I have an excellent day job and make excellent money doing it. I'm just a recreational high roller that would like to get the comps and enjoy myself without getting clobbered at the tables. If I don't bet the $100 unit minimum, the comps aren't going to be there (I know from years of playing green chips and getting SQUAT).

    I know I'm not making sense in your professional blackjack world that I don't want to optimize my bet schedule for growth of my bankroll, or divide it into the proper units ($87) and spread 1-10 or more. Sorry, I don't need the money to live. However, if I'm going to lose it ALL 1 out of every 5 trips I go on--I might as well invest it in the next "hot stock tip" that comes over the phone.

    You've helped me understand the issues with how to go about what I am looking to do. That is all you can do, and your advice has been helpful. Thanks.

  12. #12
    Don Schlesinger
    Guest

    Don Schlesinger: Re: Understood, thanks

    > I know I'm not making sense in your professional
    > blackjack world that I don't want to optimize my bet
    > schedule for growth of my bankroll, or divide it into
    > the proper units ($87) and spread 1-10 or more. Sorry,
    > I don't need the money to live.

    That doesn't mean you shouldn't bet properly. One has nothing to do with the other. We're beginning to talk at cross purposes. The comps you receive will hardly be worth the lost e.v. you have from a poor bet schedule or too large ROR. You're being penny-wise and pound foolish, in my view.

    > However, if I'm going
    > to lose it ALL 1 out of every 5 trips I go on--I might
    > as well invest it in the next "hot stock
    > tip" that comes over the phone.

    Are you sure that the ROR you're referring to is trip ROR and not long-term? If long-term is 20%, that doesn't mean you'll lose it once out of every five trips.

    > You've helped me understand the issues with how to go
    > about what I am looking to do. That is all you can do,
    > and your advice has been helpful. Thanks.

    No, I can probably do more, but not with incomplete information.

    Don

  13. #13
    Wolverine
    Guest

    Wolverine: Lurkers

    > Are you sure that the ROR you're referring to is trip
    > ROR and not long-term? If long-term is 20%, that
    > doesn't mean you'll lose it once out of every five
    > trips.

    Don, my trip and total bankroll are the same. So, yes, I think a 20% RoR means I have lost ALL my blackjack money and I have to replenish my entire bankroll if I want to go play more. Too much risk considering I haven't had a winning trip since I started using index play--and my CV scores indicate I know my indices--so mistakes aren't a major issue. At least when I bet $25 for bad, $50 for good and played BS in the 90s, I broke even nearly every trip on a 5K bankroll.

    > No, I can probably do more, but not with incomplete
    > information.
    > Don

    If I felt safer that I was telling ONLY you the information, then I would. However, if you have noticed, this thread is being followed by a few people besides you and me. That's fine--maybe they are trying to learn something too, or are laughing hysterically at my ignorance. Either way--I've let out enough information in this thread to get me into some trouble and pigeon hole my stores--and since I've never had a stitch of problems in a solid 10 years in Vegas---I'm not about to start now. Read back through the thread---I believe I've left all the cookie crumbs you need to look at the entire situation. Trust me. Thanks.


Page 1 of 2 12 LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

About Blackjack: The Forum

BJTF is an advantage player site based on the principles of comity. That is, civil and considerate behavior for the mutual benefit of all involved. The goal of advantage play is the legal extraction of funds from gaming establishments by gaining a mathematic advantage and developing the skills required to use that advantage. To maximize our success, it is important to understand that we are all on the same side. Personal conflicts simply get in the way of our goals.