How to test a strategy?

I've been trying to find a way to evaluate my playing strategy (no, not a "betting system") over large numbers of rounds, say several thousand or more. It's a pretty simple strategy, just involving don't pass and don't pass odds. Of course, the house edge remains as always, but I'd like a clearer picture of how it works over time, the volatility, variances, etc.

For many years I was the resident Excel hacker at my job, but I can't for the life of me create it in Excel. Then there's an old software program called WinCraps that's supposed to do this, but in all humility, I can't figure out how to make it work. I've tried just keeping records of plays and wins from online games (the Wizard's craps game is wonderful), but it's just too tedious. Can anyone help me out?

Thanks, see you at the tables,

Ruark

What's your strategy trying to accomplish? I imagine if you tell us, we can give you a quick analysis.

Or you can write up a quick Java simulation.

Quote: Ruark

Then there's an old software program called WinCraps that's supposed to do this, but in all humility, I can't figure out how to make it work.

Come on, man.
Your first post here is a dud.

Not enough information given.
There are NO mind readers in this forum.
All of those are over at the Craps Forum.
added:
http://www.crapsforum.com/


Now,
WinCraps version 5.1b is current.
I know a 10 year old that can program WinCraps to run simulations.
He says it is boring.
edit: he says he is 11.
my mistake. Excuse me!

Anyone, if they take a little time to read and look can too.
Steen included the best Craps Help section in it, it is free to download and read.
Examples on how to code simulations are also included
and free auto-bet files on the website.

What more do you want?

I see, you might just want someone to hold your hand and do all the work for you?
Excellent if it works!

added: the auto-bet file that is FREE and works for the pass line is called "riskofruin.bet"
It is very simple to change every reference of pass line to don't pass line.
One still has to read the instructions at the beginning and follow them, otherwise your WinCraps experience will be a disappointing one.
(you also need to have the registered version to run the hyper-drive mode)

Excel is way easy if you know how to use the included functions.
But you would like me to give my Excel workbook to you because???
(make it good, a friend of mine just died today)

Even easier is to learn the EV (expected value),
HE (house edge) and
Variance calculations, the proper calculations,
of the bets you make and you can gain lots if info just from that.
Oh, I forgot, you kept all that info to yourself.

Beware, over 98.6% (+/-) of any and all assistance offered at this forum is error filled... meaning you need to be able to prove everything offered to you, unless you pay $$$ for it and get a money back guarantee.

You are now placing your life into others hands

Good Luck


last added: sunday nfl time with South Park
OP, If you would like to have my Excel workbook without giving me a reason why, let me know that too.
You want .xls or can you handle .xlsb?
your claim... "I was the resident Excel hacker"
here is a snap.
Maybe you can make out the code
C2: 1
C3: =IF(OR(E2={"Pass","Miss","NMiss","NPass"}),1,C2+1)

I would rather not link to it
so people like Tim Allen or Jay Leno can not have it.

One bit of advice for the newbie: Do not, ever, listen to 7craps.

Seriously, I'd just do it directly, or tell us what your strategy is.

Why do you do this? Were you abused as a child? What happened to you?

Quote: 7craps

Come on, man.
Your first post here is a dud.

Not enough information given.
There are NO mind readers in this forum. All of those are over at the Craps Forum.

Now,
WinCraps version 5.1b is current.
I know a 10 year old that can program WinCraps to run simulations.
He says it is boring.
edit: he says he is 11.
my mistake. Excuse me!

Anyone, if they take a little time to read and look can too.
Steen included the best Craps Help section in it, it is free to download and read.
Examples on how to code simulations are also included
and free auto-bet files on the website.

What more do you want?

I see, you might just want someone to hold your hand and do all the work for you?

Excellent if it works!

Excel is way easy if you know how to use the included functions.
But you would like me to give my Excel workbook to you because???
(make it good, a friend of mine just died today)

Even easier is to learn the EV (expected value),
HE (house edge) and
Variance calculations, the proper calculations,
of the bets you make and you can gain lots if info just from that.
Oh, I forgot, you kept all that info to yourself.

Beware, over 98.6% (+/-) of any and all assistance offered at this forum is error filled... meaning you need to be able to prove everything offered to you, unless you pay $$$ for it and get a money back guarantee.

You are now placing your life into others hands

Good Luck

I couldn't find anything to do this type of thing, and I ended up writing some code in Perl and spitting out data that gets sent to gnuplot so I can chart various betting strategies.

Here's a chart on random data showing a strategy that bets the 4 and 10 with various amounts of pressure according to a press schedule.



I used this web page to take a sample for a finance chart so that I could plot out the money on the rail, which is what is shown.

http://gnuplot.sourceforge.net/demo/

You could also plot out the money on the felt if you wanted and anything else you want.

I recommend that if you have the skills to do so, whether you use someone else's software or write your own, you can create data in text files and chart them out with gnuplot.

I hope this helps!

Quote: Ruark

I've been trying to find a way to evaluate my playing strategy (no, not a "betting system") over large numbers of rounds, say several thousand or more. It's a pretty simple strategy, just involving don't pass and don't pass odds. Of course, the house edge remains as always, but I'd like a clearer picture of how it works over time, the volatility, variances, etc.

For many years I was the resident Excel hacker at my job, but I can't for the life of me create it in Excel. Then there's an old software program called WinCraps that's supposed to do this, but in all humility, I can't figure out how to make it work. I've tried just keeping records of plays and wins from online games (the Wizard's craps game is wonderful), but it's just too tedious. Can anyone help me out?

Thanks, see you at the tables,

Ruark

This, by the way is a chart of a betting strategy that presses into a low edge bet that starts at $25 or more and adds pressure. 7 craps, when I began a discussion about this strategy that made the claim that the house edge for this combination of bets creating a new bet, made some similarly humiliating comments about what I was doing without truly understanding what he was talking about. Even now, I'm not even sure he understands what it is that I am doing, but that's alright: he like to humiliate people more than to try to help them.

Don't take it personally. This is 7 craps' problem, not your problem. He's obeying all the rules and I think he just likes the attention that he gets doing this type of thing.

Here's the various schedule for the pressure that I'm applying. All my limits are set to 1000 regardless of this schedule in the above graphs.

The pass line with no odds is shown to compare to a 1.41% house edge with minimum volatility on a $5 bet.

@buy_press_schedule7 = ( 5, 15, 40, 120, 360, 1080, 3240 );
@buy_press_schedule6 = ( 5, 15, 25, 75, 225, 675, 2025 );
@buy_press_schedule5 = ( 5, 15, 30, 80, 260, 750, 2000 );
@buy_press_schedule4 = ( 5, 10, 25, 60, 160, 460, 1350 );
@buy_press_schedule3 = ( 5, 10, 20, 55, 160, 420, 1200 );
@buy_press_schedule2 = ( 5, 10, 25, 50, 75, 100, 150, 250, 500 );
@buy_press_schedule1 = ( 5, 10, 15, 20, 25, 40, 80, 100, 150, 250, 300, 350, 400, 450, 500 );

" but that's alright: he like to humiliate people more than to try to help them.

Don't take it personally. "

Every time somebody tells me " Don't take it personally" I take it personal !

Come on ahigh, I thought better of you.

Quote: buzzpaff

" but that's alright: he like to humiliate people more than to try to help them.

Don't take it personally. "

Every time somebody tells me " Don't take it personally" I take it personal !

Come on ahigh, I thought better of you.

Buzz on the other hand is right about everything that he says. All of his comments reflect an in depth knowledge of everything true and correct. Do not doubt buzz. All his criticisms are entirely valid.

Ahigh, you missed my meaning entirely. I admire that you usually state your opinions are are always willing to defend them.

It's just that I never expected you to hide behind that weasel expression " Don't take it personal "

Hey, I am probably wrong more often than not. But the only opinion I have is mine at present.

That can definitely be changed in dialogue. It's what a forum is supposed to be all about.

I will refrain from responding to your posting in the future ,if that makes you happy.

HAVE FUN !

Quote: buzzpaff

It's just that I never expected you to hide behind that weasel expression...

expression?
I found this one.

Quote: Ruark

I've been trying to find a way to evaluate my playing strategy (no, not a "betting system") over large numbers of rounds, say several thousand or more. It's a pretty simple strategy, just involving don't pass and don't pass odds. Of course, the house edge remains as always, but I'd like a clearer picture of how it works over time, the volatility, variances, etc.

For many years I was the resident Excel hacker at my job, but I can't for the life of me create it in Excel. Then there's an old software program called WinCraps that's supposed to do this, but in all humility, I can't figure out how to make it work. I've tried just keeping records of plays and wins from online games (the Wizard's craps game is wonderful), but it's just too tedious. Can anyone help me out?

Thanks, see you at the tables,

Ruark

Ruark,

If you can describe your strategy to me in plain english, I can code it up and apply your strategy to some random rolls. Additionally, if you provide me with roll data, I can run your strategy against your rolls as well. I'm always interested to add play strategies to my software, and unless your strategy is something you would rather not share, I'm happy to help you out.

I know that there are many math people on here, including 7 craps, that can give you scalar numbers and percentages, but for me, I really enjoy seeing the shapes of the graphs in trying to identify a curve pointed upwards that I think I might have a chance of catching the updraft on. It's probably due to the fact that I do a lot of investing and I just like looking at charts. Anyway best of luck to you, even if 7 craps ends up being more helpful.

Thanks, Ahigh and the rest of you for all the posts. This was indeed my first post to this forum. I've also been on crapsforum.com a few times; I got to this forum via the wizard's website.

Myself, I'm 61, been to Vegas many times, but I'm not a big time gambler. Much of my experience comes from my aunt, who was a Pit Boss in the Reno area casinos for 35 years, usually BJ. I'm pretty much a realist when it comes to gambling. I know you WILL lose your money over time, no matter what you play, with the possible exception of card counting, which is becoming almost useless these days. I do NOT believe in "hot" or "cold" tables, controlling dice, progressions, etc. Much to the dismay of casino managers, I'm sometimes seen putting $10 in a penny slot and playing for a couple of hours at 1 or 2 cents per roll, just relaxing and enjoying myself (and tipping the bar girls generously), but never for "serious money." Once in a blue moon I might make a silly horn bet, just for fun. I've even been known to throw a few chips onto a roulette table, but again never "seriously." I haven't played BJ in decades, since I played it in Reno back when you could find 3:2 single-deck with $1 minimums. Boy, those were the days..(-;

Anyway, that being said, my usual craps playing strategy seems to work pretty well most of the time. In fact, I've played hundreds, possibly thousands, of rounds on the online sims (mostly Wizard's) and I honestly can't remember ever not being ahead at SOME point, although statistically, as always, ruin is inevitable. It has quite a bit of volatility, so basically what I do is ride the roller coaster until I'm ahead, then quit.

There are two caveats:

First, it requires a table with at least the 3x-4x-5x odds rule. This allows you to take up to 6X odds on the DP points.

Second, it requires some grinding; you might have to sit there through a dozen or more rolls, waiting for a 7, much to the chagrin of the boys. For this reason, I like to play tables alone or with just 1 or 2 other players. When it's packed shoulder to shoulder with screaming yuppies, play is so slow it's just not practical to grind. Not to be a wet blanket or anything, but I don't gamble to "party."

One thing I'd like to do is compute all the high points my bankroll reaches before going negative. I'd like to find an "average high" that would help determine a quitting point, something I don't really have yet. I'd also like to be able to analyze the rise and fall of the bankroll from one round to the next (e.g. a line graph), and look at its trend line, that sort of thing.

It's nothing fancy.

1. Bet one unit on DP.
2. Come-out roll. The usual rules: 7/11 loses, 2/3 wins, 12 stays.
3. If there's a point, lay odds as follows:

4 or 10: 5X
5 or 9: 4X
6 or 8: 3X

That's it. I just grind it out, and it seems like I always eventually come out ahead, although the negative swings can be terrifying at times. That's why I'd like to see some long-term simulation, which I just can't duplicate. I KNOW the house edge has to be in there somewhere. I'd also like to try it at 6X-4X-2X, just to compare the trend lines.

Thanks again, everybody have a great day.

Quote: Ruark

That's it.
I just grind it out, and it seems like I always eventually come out ahead, although the negative swings can be terrifying at times. That's why I'd like to see some long-term simulation, which I just can't duplicate.

This is so easily done in WinCraps it is not funny.
WC is just $19.95. Most piss that away every week.
Learn to do this yourself.
What is the fear??

Plus...
You still do not supply needed parameters.

What is your Starting bankroll?
(again NO mind readers here, I could say it is $500,no, I mean $300, but you would say it is $1000)

flat bet size? $5 or $10 or...

how long of a session are you willing to play, or is this just for one lifetime long session?

The win goal.
I assume you want to see the actual distribution of all win goals, I would, to make your conclusions.

You may be very surprised at the actual number of times one will NEVER be ahead making line bets with odds in Craps.
It is because of the variance.

1 million sessions
Trying to win just $5

$500 bankroll $5 bets for your DPass and 345X odds
Bankroll was busted . . = 3.852% of the time ( 38523)
Win goal was met . . . = 96.148% of the time ( 961477)

1 million sessions
Trying to win just $5
$1000 bankroll $5 bets for your DPass and 345X odds
Bankroll was busted . . = 2.082% of the time ( 20818)
Win goal was met . . . = 97.918% of the time ( 979182)
2% is 1 in 50.
Expect, on average, 1 in 50 to bust.
It could also be 2 in 50. Damn.

1 million sessions
Trying to win $10
$1000 bankroll $5 bets for your DPass and 345X odds
Bankroll was busted . . = 2.423% of the time ( 24232)
Win goal was met . . . = 97.577% of the time ( 975768)

Notice the bust rate going UP as the win goal goes UP.

(above sim data from free program, at the bottom of the page, found here
http://members.dslextreme.com/users/stevefry/gamblsim.htm
Easy to convert to the language of your choice, I did it in Basic
Way faster than WinCraps for some basic sims, but lacks a lot of data output in the free program)

It sounds to me as though you don't like to leave the table behind, and are leaving most of the time with relatively small wins compared to your buy-in; in other words, when you do lose, it will be after chasing your losses quite far, and consequently you'll lose a lot more than you've been winning. That's where the house edge is. This strategy has an SD/EV of about 200, so you're probably going to see some pretty dramatic swings at any session, but you can't rely on them.

EDIT: Dammit.

7craps:

Thanks for taking the time to reply, I appreciate it.

It wasn't clear on the WinCraps demo that HyperDrive wasn't working. I'll probably buy it, then.

Starting bankroll: 1,000 units. I've never really figured this into it.

Bet size, as in the description of the strategy, is 1 unit.

I hope you kept in mind that the odds are 543x, not the usual 345x.

Your simulation really doesn't reflect what I'm doing, because you're playing it with a specific win goal, e.g. $5. It assumes I will quit when I reach it, then start over. I haven't been doing that, and I very frequently get ahead 40 or 50 units.

Quote: 24Bingo

..are leaving most of the time with relatively small wins compared to your buy-in; in other words, when you do lose, it will be after chasing your losses quite far, and consequently you'll lose a lot more than you've been winning. That's where the house edge is. This strategy has an SD/EV of about 275, so you're probably going to see some pretty dramatic swings at any session, but you can't rely on them.

Right now I'm just playing with units. Being ahead 20 units is pretty significant when one unit equals 50 or 100 dollars.

But what you're saying, if I understand correctly, is that EVENTUALLY I'll have a bad run that will take the 1,000-unit bankroll down to 0, and when that finally happens, however long it may take, it will cancel out all the wins.

If that's how far you're willing to chase your session losses, yes, eventually you'll lose everything. How far you're willing to chase your losses will determine both how unlikely you are to lose, and how much you lose when you do. What I guarantee, however, is that those losses will cover the wins, plus a bit.

Quote: Ruark

....because you're playing it with a specific win goal, e.g. $5. It assumes I will quit when I reach it, then start over. I haven't been doing that, and I very frequently get ahead 40 or 50 units.

You never said your starting bankroll.
Your starting bankroll decides when you start over.

So I picked 2 of them starting bankrolls.

I played until a profit of at least $5 (because with lay odds most times the $5 was overshot)
was hit or you busted your bankroll whatever came FIRST.

Yeah, I see you do not believe one can bust $1000 bankrolls trying to win at least $5 or bust $1K trying.


But of course NOW you state having a 1000 unit bankroll.
so betting $5 flat bets would make that a $5000 bankroll.
betting $10 flat bets would make that a $10000 bankroll... etc.

The lay odds are correct for a 345X table. 6x lay for each number.
$5 flat bet and always a max $30 on any point number.

Easy sim for me.
bust a 1000unit bankroll or Win at least 1unit whatever comes first.
I want to see the busts.

Bankroll was busted . . = 0.632% of the time ( 6320)
Win goal was met . . . = 99.368% of the time ( 993680)
about 1 in 159 sessions, on average, bust a 1000unit bankroll just trying to win at least 1unit

I sure hope your simulations show you win more than 1000units the other 158 sessions on average.

Mine show they do not most times on average.
The only thing in your favor to show a profit after X number of bets is the odds bet, it adds FREE variance.
But you will not believe my data even if I gave you all the code, my computer and $10,000US.
You need to prove these things to yourself.

You can do the math from here.

ev and sd tells the rest of the story, ask someone to calculate those for you and use the central limit theorem for your solution.
or just program WinCraps and run the sims yourself and set up all the charts in Excel to make you happy.
It is very easy to do.

Quote: 7craps

I played until a profit of at least $5 (because with lay odds most times the $5 was overshot)
was hit or you busted your bankroll whatever came FIRST.

The lay odds are correct for a 345X table. 6x lay for each number.
$5 flat bet and always a max $30 on any point number.

Easy sim for me.
bust a 1000unit bankroll or Win at least 1unit whatever comes first.
I want to see the busts.

Bankroll was busted . . = 0.632% of the time ( 6320)
Win goal was met . . . = 99.368% of the time ( 993680)
about 1 in 159 sessions, on average, bust a 1000unit bankroll just trying to win at least 1unit

I think you're playing the odds the way I indicated. 5 units on 4/10, 4 units on 5/9, 3 units on 6/8.

The main thing you're doing that I'm not doing is stopping when you hit the win goal. I have NOT YET, just "manually" playing simulators, been able to bust, or even come close. The lowest I've gone (before heading back upwards) is 952, and then I eventually got back up over 1,000, frequently going up to 1040-1050 before having to turn it off and go to bed or something. It feels, just subjectively, like this would be a pretty nice strategy (no, not a "betting system") if one could develop effective win/loss limits.

I'd like to approach it from the standpoint of finding out how many rounds it takes to bring the bankroll down to 0. Then look at the range, SD and some other statistics.

One problem you might run into is that they might not let you lay those exact odds, especially the 4x on 5 or 9.

Quote: Ruark

I've been trying to find a way to evaluate my playing strategy (no, not a "betting system") over large numbers of rounds, say several thousand or more. It's a pretty simple strategy, just involving don't pass and don't pass odds. Of course, the house edge remains as always, but I'd like a clearer picture of how it works over time, the volatility, variances, etc.

For many years I was the resident Excel hacker at my job, but I can't for the life of me create it in Excel. Then there's an old software program called WinCraps that's supposed to do this, but in all humility, I can't figure out how to make it work. I've tried just keeping records of plays and wins from online games (the Wizard's craps game is wonderful), but it's just too tedious. Can anyone help me out?

Thanks, see you at the tables,

Ruark

get AW! Craps for you ipad/iphone. It doesn't keep stats, but you can play for hours...it's fun.