***Official HEM Red Line Discussion Thread***

[dybboss]

i'd say that hem2 is wrong, cose it is 'open beta' xd
anyway just have a look at hands, maybe you'll se where it is mistaken, I am not gonna install till you guys 'beta test' it for 6 months ;p

[wayneking7]

^ pretty sure at the rate things are getting fixed in HM2 u will need to wait 2 years with the rest of us

[ymu]

Originally Posted by vookenmeister

I've been running cold in the high dollar DONs so I decided to review the numbers behind one of my failed $50 DONs.
In it I actually folded AQ shortly before the four hands you so below. I folded AQ because there were several shorties and the big stack was calling off
or shoving against any raise. it was a standard fold for many of you that read this thread...

In the four hands below, I actually sit on the rail semi-safe in chips while i watch all four shorties win an all in against the big stack.
I ended up losing later when i ran my A10 into AK (of course, standard shove though with monster blinds).

What ate me up is that it showed my EV from the tourney to be lower than I expected because it mostly captured the one hand where i ran A10 into AK but
didn't give any credit for those four times when the tourney almost ended and I would've won $100.

THis thread is so long. I read thru it years ago and don't recall seeing a good explanation. If there's one in here please just point me
to it, quickly summarize, tell me I'm an idiot or whatever.

I've been puzzling over this one, and I read the thread not so long ago, and I'm not sure either.

This is how I *think* you need to interpret the redline ...

1. To get very unlucky, you have to get it all in way ahead and to get very lucky you have to get it all in way behind. Coinflips can't shift the red line much.

2. It only looks at luck on all-ins before the river. The $EV attached to the hand is calculated from the ICM value of the pot multiplied by your chance of winning it. You win, you got lucky (even as a favourite). Getting lucky/unlucky on the bubble and ITM will have a bigger $EV impact than if it were in the first hand or when you called a shortstack push.

It counts these Sklansky-ICM dollars up over the course of a tournament and presents it as a running tally. The number isn't directly comparable to your actual winrate - it's often much more than you could have won/did lose. Hence the potential for the redline to be many buy-ins away from actual - it's multiple-counting a lot of equity in the same tournament.

The theory is that the red-line will always catch up with your actual winnings eventually, and might follow them quite closely. That's because ICM is a decent model of your predicted tournament winnings. The calculations for each individual tournament are just a way of keeping a running tally on luck in terms of EV rather than modelling an actual tournament result.

I do wonder if aggressive STT players are more likely to run some way above their adjusted line, simply because are more likely to be behind when called, but they've accumulated enough chips to recover so the impact on their $EV is more limited (running into those ICM limitations here: the impact of skill plus overvaluing shortstacks).

[vookenmeister]

agreed.

below is another thing I realized recently that makes my numbers look skewy and throws me off. I posted this is in a DoN thread but haven't seen an answer yet. This is in regards to using HEM on the Merge Network to track DoNs...

---

Anybody solve the issue with HEM EV and DoNs??

SPecifically this... using Hold Em Manager V1 latest update (and previous updates). the all in EV adjust luck calculations are done based on a top 3 SNG finish payout and not the shared double up top 5 calculation. If you use HEM and don't know what I'm talking about, replay a hand from a DoN and click on ICM. You'll see it lists the payouts as the standard, 50%, 30%, 20% top 3 finishers and not the DoN format. Then if you run the math, you'll see you're EV calcs are based on that.

Yes, I know EV isn't perfect. Yes, I know there is a red line thread (and no answers there). Yes, I know there is a HEM forum on their product site (no good answer there. it appears this was fixed in Stars).

I didn't use to care. but lately I have been playing a lot more consecutive tables so it is very hard for me to analyze my play as I go. I would like to rely on the red line to get a feel on how I'm running. and of course, since my red line is above my green line since joining Lock I want to comfort myself by saying I should be winning more. However, even this limited analysis of luck is skewed in some fashion since the payouts are wrong.

seriously though... has anybody who plays DoNs noticed this? anybody heard of a fix? anybody care?

[ymu]

You might be able to edit the tournament payout structure somewhere. Look under tags, or ask on the HM forums.

You might be better off having a look at the races report and checking out your equity on all-ins. Filter it by hands where you called an AI and make sure you're getting the right sort of (average) ICM equity in those spots. Filter it by hands where you lost an AI and see how often it busted you. That sort of thing.

I don't think the red line tells you anything much about your tournament play because so much of the equity you have before each AI depends on how you played earlier in the tournament.

[vookenmeister]

Originally Posted by ymu

You might be able to edit the tournament payout structure somewhere. Look under tags, or ask on the HM forums.

You might be better off having a look at the races report and checking out your equity on all-ins. Filter it by hands where you called an AI and make sure you're getting the right sort of (average) ICM equity in those spots. Filter it by hands where you lost an AI and see how often it busted you. That sort of thing.

thx... will check it out. I didn't see anything on HM forum. will post there next if I can't figure it out.

[imluckynotgood]

Ok so I've been skimming this thread (I admit I could have missed a portion that answered my question), but I play 6max super turbo STTs on Carbon and after an albeit small sample size my profit is WAY above my luck adjusted winnings red line. Practically mirror images!

My method is play for first in each tournament, pushing strong hands 6 handed, and either busting in 6th or making it into 3 handed-ish ~3:1 chip leader. At this point I aggressively steal blinds from smaller nit fish stacks until someone finally pushes with the strong hand they've waited on and i call with any two (they have half their stack in blind already) and can't see giving them a walk. Is this completely flawed play on my behalf and should I expect my BR to implode inevitably? Thanks for any insight and please ask questions if I need to specify anything..

Ex3 handed) Villian pushes remaining t600.
I have t300 in blind, with t2400 behind, so I call off the remaining t300 with 46o

I assume this destroys my EV when he turns over A6..

[vookenmeister]

Strat sounds fine and Standard call. If your sample is small and u r crushing then your red line must be lower. U might be playing great but r also running well. How big is sample?

[imluckynotgood]

an embarrassing 220.. I just figured I'd get to the bottom of this before I wasted my time if I'm really that lucky and bad. lol

[Max Cut]

imluckynotgood, you are not going to perpetually separate your red line and green line through any style of play.

[imluckynotgood]

Max I understand what you're saying as anyone's EV is a more 'true' picture of their play, but I'm wondering if all these "blind levels relative to stack size" decisions I'm making are taken into context in this red line.

i.e. He could show me his AK, but I'm going to call off 25% of my stack with Q3 bc im not folding and letting him double up by stealing the blinds.

[Max Cut]

Yes, they are (every decision is). But your number of games is way too small to tell anything from it at all. For those games, you want like at least 10,000 before you try to evaluate your play based on the red line. I personally think a lot more than that is needed to have any reasonable confidence.

[ymu]

It is too small to say, but I do think that it's possible for a skilled aggressive STT player to have actual winnings persistently above the red line.

This is because the red line only measures what happens at points in the tournament where the $ value of your chipstack might potentially be realised (usually to zero, unless it's on or after the bubble).

If you have an above average stack when it gets to push/fold time for most players, then you have a higher $EV to lose (bigger red line swings) and a higher chance of making the money than your opponents. Players who steal aggressively will often get caught with the worst hand, but they will survive most AI confrontations because they have a deeper than average stack, and they will steal back into the money a good proportion of the time when they lose the AI.

Over the very long run their luck will average out, but their winnings could still remain substantially higher because of the skill edge, which ICM ignores.

This only applies to tournaments, of course. The red line is much easier to interpret for cash.

I know there's been some discussion about alternative ways to present the redline data for tournaments - there's a thread somewhere on the Holdem Manager forums where they seek comments. One idea is capping it at the max winnings or a lost buy-in to make the scale more realistic - but this introduces bias so the red line is no longer a good measure of $EV.

Has anyone looked at apportioning the prize pool according to total $EV earned in the course of a tournament? The average $EV per player must equal the prize pool, so would giving them their $EV/total prize pool as the adjusted winnings work at all?

[vookenmeister]

Originally Posted by vookenmeister

thx... will check it out. I didn't see anything on HM forum. will post there next if I can't figure it out.

Yum ->. Where do I find the races report you referred to?

Your theory makes sense to me. I need to break my reliance on the red line as a comfort that I should be doing better. It has too many flaws that make it hard to distinguish all in luck. If I could get my green line close it I promise not to looks (lol)

Seriously. I think checking my all ins would help. Concerned I am bubble shove happy

[ymu]

It's listed under more reports on the ribbon under the main Reports tab.

It's more important that your shoves were +$EV than that you got it in ahead. Sometimes it's just luck whether you've won enough shoves to survive a losing AI.