A SAGE mistake / sng's

[jay_shark]

I noticed a small mistake in the numbers provided from SAGE .
It says that for R=1 ; that is , after the blinds have been taken , and that you're left with 1 big blind , then you should jam with 6-2 suited or better . It should read 6-2 offsuit or better since 6-2 offsuit wins slightly over one-third of the time . This means that you should fold with only two hands ; namely , 2-3 off-suit and 4-2 off-suit .

We're in Nash equilibrium if we're satisfied with our strategy given what the other player has done . We would not be satisfied if we had folded 6-2 off-suit since we're getting 2:1 (Villain calls 100% of the time) and that 6-2 off-suit wins slightly more than 34% of the time .

[roggles]

According to the table I use we shouldn't even be jamming with 6-2 suited. 62s has 16 points, but we should have 17.

[jay_shark]

There have been several mistakes pointed out by others on when you should jam/fold according to their Power Index numbers . There have been many here who have jammed incorrectly according to their number when they should have folded . Your best bet is to memorize the hands superior to the cut-off point or accept the fact that the Power Index numbers can be misleading .

[daveT]

"A SAGE mistake/sng's"

[ QUOTE ]
using it

[/ QUOTE ]

Edited by daveT.

<------loc.

[HajiShirazu]

Why do people follow this when you can easily just use sng wizard quiz and play substantially better?

[jool]

I was going to ask the same thing. Seems like ICM is a much better tool than SAGE for exactly the reason people are objecting to SAGE- ability to take reads into account. It is still a "system", perhaps, but one that will much more accurately inform real-game situations. Am I missing something?

[jay_shark]

There is no such thing as ICM in heads up sng's .

SAGE is nevertheless very effective towards the end of a game ; especially against a strong opponent . However , with experience , you may learn to deviate slightly from game theory strategy to exploit the mistakes made from your opponent's .

[jool]

it's been years since my days in stt, but isn't icm just assigning chip values based on your opponents range, and seeing how much real $ that makes various actions worth? i can understand how the payout structures would make one facet of that a non-factor, but i don't get how it's not applicable at all. what am i misunderstanding?

[daveT]

On SNG Wizard:

I don't see how it applies to HUSNGs.

[jool]

Does sngwiz/sngpt automatically assume 50/30/20 payout? If not couldn't you set a 100/0 payout, put one opponent, and have a much better tool than SAGE?

If not, how hard would it be for someone familiar w/ icm to make a program that adjusts to the 100/0 payout of a headsup sng? I feel like I must be missing something, cuz why didn't someone else notice, but can't think of what it is.