PDA

View Full Version : "new" statistic "idea", possibly worthwhile?


citanul
11-23-2005, 03:01 PM
so we've got ROI

ROI = (winnings - investment)/investment

and that's cool and all that.

my only problem with it is that sometimes it's negative, and there's certain calculations i want to do (or more accurately, tell others to do) that aren't schway with the whole negative thing.

namely, if one were to have a set of predicted ROIs for the steps, and have a negative ROI at the top couple steps, you couldn't multiply through your ROIs for an overall ROI.

however, you could use "total return on dollars" or whatever the right term is.

that would be, in layman's terms, 1+ROI.

this gives you, when you multiply it by your buyin, the total return per tournament.

which means that it's a per dollar amount you're returned.

which means that if you have a bunch of them (some possibly lower than 1) you can multiply them through to get an overall number.

so yeah, this isn't new or anything, but i've told a few people lately to multiply through their step ROIs (theoretical) to see what kind of stats they'd need to rack up to make various amounts of money, and i thought that it would be nice to show what i meant while i was at it.

So, a review:

A player who has a TR (total return) = 1.1 has an ROI of .1, and if they were playing say a $1000 buyin tournament, they expect to end up with $1100 at the end on average. A player with a TR of .9 in the same spot would expect to get up with $900 on average.

note, this doesn't mean that if a player has a TR < 1 at a higher step they are necessarilly a -ev player at the steps in general, it means exactly what i said above, that there's an easy way to compute chain expectations, and it will give you an expected total return on dollars invested, which of course will play nicely with whatever spot you decide to buy in at. playing around with these statistics enough would allow you to make nice decisions like what step level makes your hourly earn the highest (mostly important for players who expect themselves to be -ev at top steps), and create spreadsheets capable of telling you information on net investments starting at varying points in the buyin scale.

c

The Yugoslavian
11-23-2005, 03:03 PM
W00t?

Yugoslav

11-23-2005, 03:20 PM
I've been thinking about playing the ministeps and I've been using "TR" to estimate my total ROI and such.

P.S. I don't want to start a new thread about this. What BR do I need for the ministeps starting from step 2? step 1?

Insty
11-23-2005, 03:22 PM
So you are proposing Total Return (TR) = 1 + ROI?

This seems like a simple calculation for any further calculations you want to do, that you can calculate if you know the ROI anyway without making other things more complicated than they need be.

I'm not sure this would be generally useful.

citanul
11-23-2005, 03:24 PM
[ QUOTE ]
P.S. I don't want to start a new thread about this. What BR do I need for the ministeps starting from step 2? step 1?

[/ QUOTE ]

well, offhand, i don't remember what the buyins are /images/graemlins/smile.gif or the payout structures! /images/graemlins/smile.gif but using whatever statistics you've been using, you should be able to come up with some approximate numbers for a semi-safe bankroll. i think that someone a while back posted a steps simulator, and if not, you could do something where you make up finish distribution information for each step, start with 100 at say step 1 and step 2, and figure out your "expected" results by iteration (i swear it really won't take as long as it sounds, probably only a few minutes. i'm sadly not equipped to even hazard a guess without those first couple of things up there (and wouldn't have time to sit down and do the maths for you if i did /images/graemlins/frown.gif

c

11-23-2005, 03:26 PM
[ QUOTE ]
[ QUOTE ]
P.S. I don't want to start a new thread about this. What BR do I need for the ministeps starting from step 2? step 1?

[/ QUOTE ]

well, offhand, i don't remember what the buyins are /images/graemlins/smile.gif or the payout structures! /images/graemlins/smile.gif but using whatever statistics you've been using, you should be able to come up with some approximate numbers for a semi-safe bankroll. i think that someone a while back posted a steps simulator, and if not, you could do something where you make up finish distribution information for each step, start with 100 at say step 1 and step 2, and figure out your "expected" results by iteration (i swear it really won't take as long as it sounds, probably only a few minutes. i'm sadly not equipped to even hazard a guess without those first couple of things up there (and wouldn't have time to sit down and do the maths for you if i did /images/graemlins/frown.gif

c

[/ QUOTE ]

I know.. but I'm lazy and left my TI-89 at school. Oh well; I'll break out Excel. Thanks anyway.

skipperbob
11-23-2005, 03:32 PM
You picked a helluva week to start sniffing glue again /images/graemlins/tongue.gif

citanul
11-23-2005, 03:36 PM
sorry brah, fortunately, excel pwn3s