A S&G Parlay Strategy

[Daxonovitch]

Better to just ask here, email is for personal people. [img]/images/graemlins/smile.gif[/img]

Besides, the more intelligent conversation we have, the more other people can benefit from it.. If you have questions, maybe someone else does too, and maybe someone else will find unique viewpoints to the scenarios and contribute.

[fiyah]

I was just going to ask you what your style of play is for the upper levels since you say you do pretty well. For instance are you pushing all in frequently, etc.

[Daxonovitch]

That's the name of the game, now, isn't it? [img]/images/graemlins/smile.gif[/img]

In general, I find that aggression and the gap concept will take you far. At the lower levels though (30 and down), you have to moderate your gap hands.. For example, let's just simplify and put all the hands in a range from:


[72o-----J4o------Q7s------Ax------99------AKs-----AA]


Obviously this is a gross simplification, but for theoretical purposes it works all right. In a 30 and lower SNG you will open raise with a "tighter" group of hands (say, from Ax, maybe even A9+) and call a raise with a tighter group of hands (99+). This also assumes level 4 (50/100 blinds) and higher, as there's just too much calling going on in the early stages to do anything here. In a 50 and higher SNG you can shade down your requirements to open-raise, and provided you are playing against someone strong, perhaps even shade down your requirements to call a raise. In a high level SNG where people are playing extremely tight, I might open with any2, and bet the pot on the flop if I have enough chips and they check to me or I'm in EP. Play is simply too tight by most players.

That's it in a nutshell, obviously there's a great deal more to it - you will learn with time and practice, and an open mind dedicated towards improving your own game.

[Daxonovitch]

Posted for reference:

C:\java>java ParlayTest

* Parlay Strategy 1 *

Strategy submitted by ChrisD

Play a typical game. Let's say it was a $33. If you are above the
watermark before the game started, and you win 1st place (+$117),
play the largest level you can with those winnings without a
negative sum for the two events - in this case, a $109 would be
acceptable. This is termed the "parlay" game. In any parlay game,
should I win *1st* place, I continue the parlay to the next
highest level that I could afford. If I win 2nd or 3rd, the
parlay is over and I continue with SNGs at the "typical" level.

Running ten million iterations of parlay.............Iterations complete!
Successes: 2294056 (22.94056%), average BR: $2535
Failures: 4811192 (48.11192%), average BR: $1254
Losses: 2894752 (28.947521%), average BR: $679
Furthermore, you ROR'd your bankroll 0 times, around 0.0% of the time.

Your EV for this strategy is: <font color="red">$1381.42</font>


* Baseline Strategy '30' *

Strategy submitted by ChrisD

This strategy is the simple approach - take your bankroll and divide
by 30 - the next lowest buyin is your next game. For example, if you
have $661, you play a $22. If you have $659, you play a $11.

Running ten million iterations of baseline.............Iterations complete!
Average finishing BR: $1038
Furthermore, you ROR'd your bankroll 0 times, around 0.0% of the time.

Your EV for this strategy is: <font color="red">$1038.00</font>

[Amid Cent]

Looks like you made an error in a calculation. In the success example you posted above, you gave yourself +$1000 for a win in the $55 tournament. Unless they changed the payouts on the $55's and didn't tell me, a win is usually worth $250.

[Daxonovitch]

[ QUOTE ]
Looks like you made an error in a calculation. In the success example you posted above, you gave yourself +$1000 for a win in the $55 tournament. Unless they changed the payouts on the $55's and didn't tell me, a win is usually worth $250.

[/ QUOTE ]

Oof! Good catch.. this changes the numbers a bit.. I'll edit the posts. [img]/images/graemlins/shocked.gif[/img]

[Daxonovitch]

I'm crushed...

* Parlay Strategy 1 *

Running ten million iterations of parlay.............Iterations complete!
Successes: 563809 (5.63809%), average BR: $2441
Failures: 6144953 (61.449528%), average BR: $1127
Losses: 3291238 (32.91238%), average BR: $683
Furthermore, you ROR'd your bankroll 0 times, around 0.0% of the time.

Your EV for this strategy is: <font color="red"> $1054.95 </font>

--

You were correct, I had a bug in the software. [img]/images/graemlins/blush.gif[/img] The EV now is *much* closer to the baseline EV, but interestingly, it's still higher (by only a few bucks). However, of note is the fact that losses now constitute a 1/3 chance and the parlay is successful only 5% of the time. [img]/images/graemlins/shocked.gif[/img]

Now the question is, can we find a better parlay strategy? I'm so embarassed..

[Daxonovitch]

C:\java&gt;java ParlayTest

* Parlay Strategy 1 *

Strategy submitted by ChrisD

Play a typical game. Let's say it was a $33. If you are above the
watermark before the game started, and you win 1st place (+$117),
play the largest level you can with those winnings without a
negative sum for the two events - in this case, a $109 would be
acceptable. This is termed the "parlay" game. In any parlay game,
should I win *1st* place, I continue the parlay to the next
highest level that I could afford. If I win 2nd or 3rd, the
parlay is over and I continue with SNGs at the "typical" level.

Running ten million iterations of parlay.............Iterations complete!
Successes: 564816 (5.64816%), average BR: $2441
Failures: 6142857 (61.42857%), average BR: $1127
Losses: 3292327 (32.92327%), average BR: $683
Furthermore, you ROR'd your bankroll 0 times, around 0.0% of the time.

Your EV for this strategy is: <font color="red">1055.03</font>

* Baseline Strategy '30' *

Strategy submitted by ChrisD

This strategy is the simple approach - take your bankroll and divide
by 30 - the next lowest buyin is your next game. For example, if you
have $661, you play a $22. If you have $659, you play a $11.

Running ten million iterations of baseline.............Iterations complete!
Average finishing BR: $1037
Furthermore, you ROR'd your bankroll 0 times, around 0.0% of the time.

Your EV for this strategy is: <font color="red">1037.0 </font>

* Baseline Strategy '10' *

Strategy submitted by ChrisD

This strategy is the simple approach - take your bankroll and divide
by 10 - the next lowest buyin is your next game. For example, if you
have $221, you play a $22. If you have $219, you play a $11. This is
included here for illustration purposes, as this method is acknowledged
to be fairly risky.

Running ten million iterations of baseline.............Iterations complete!
Average finishing BR: $1515
Furthermore, you ROR'd your bankroll <font color="red">69</font> times, around 6.8999996E-4% of the time.


Your EV for this strategy is: <font color="red"> 1515.0 </font>

[AleoMagus]

My initial reactions to this are somewhat contradictory

On the one hand, I like it. I think players should allow for jumps in stakes before they really have a proper roll. I think this helps make permanent transitions easier and I think it keep us sharp becasue we tend to really think a lot when playing for high stakes and I think it also gives us better perspective on the lower stakes.

On the other hand. I don't like it. In theory, I would if you could guarantee a positive ROI at all levels, BUT for many (most?) this will not be true. What this means is that your overall ROI will be less than if you concentrate primarily on a level where you have a positive expectation. If your ROI is slightly negative at $109 or $215 (for example), then this strategy is tantamount to just going to a Baccarat table and placing a $109 or $215 bet now and then.

Regards
Brad S

[Daxonovitch]

If your ROI is negative at any level then you are a losing player and no amount of parlaying will post a +EV bet over time.

However, it is possible to tweak a strategy such that you generally play within your level, but trim off a first place win every once in a while to play in that $100 or $200 for the small chance that you cash, in which case your bankroll increases relatively dramatically.

Furthermore, the data that I've run this with are *my* data so this is not to say that this will be an accurate guess for everyone. In addition, a $1500 goal with $800 is ... fairly difficult. Basically you're looking to triple your money in a short amount of time, and no strategy will be able to accomplish this without *some* increase in variance and the corresponding decrease in ROI.

I'm in the process of writing a program that I think the S&amp;G crowd will enjoy - I'll take the statistics of anyone who wishes them (perhaps I can even release the program to 2+2), as well as several "common" strategies and post the EV, ROI and RoR numbers for each over 5 sngs, 10 sngs, 20, 50, 100, 1000, and 10000. 10,000 S&amp;Gs can reasonably be expected to be a number that not many players will achieve, even in a couple years or more of playing. If you do, though, more power to you!