Monitoring how well you do in MTTs (only) - opinions

[darthdimsky]

I'm trying to arrive at some KPIs (Key Performance Indicators) to objectively look at how well I'm doing at MTTs. Here are my thoughts:

PT4 indicators:

ITM%: While this is a good baseline it's one dimensional. A good ITM% doesn't necessarily mean you're doing well in ROI terms. Here's an article explaining this in more detail.
ROI: An ROI could also be one dimensional. For e.g. a player could bink the only mid stakes MTT he plays for a 10K payout and still not cash for the next 9998 x $1.00 MTTs he plays for a positive ROI.
Net Expected BB won: This measure is the best I've found so far in PT4. It isn't result oriented. It measures the expected outcome of your hands. While comparing against actual BBs won it clearly takes into account variance and is probably the best indicator of how you're performing. Ref first screenshot. Next Expected outcome is explained in more detail with an example here.

Non PT4 indicators:
Grouped positions: PT4 will easily display your finishes. But a 500th place finish in a player pool of 600 vs 6000 (ITM finish) is very different in it's context. One is an early finish and the other, possibly, is a middle late finish. I use pokerprolabs (ref screenshot 2) and the % breakdown is also another good indicator of how of how well you're doing with respect to your player pools. I'd love to see something like this in PT4. If anyone's seen anything like this that would be awesome.

If you're doing something with PT4 for MTTs that helps you, can you please share it in this thread?

Dims

 

[DaveE]

I don't use a tracker but I also think those early/mid/late stats are the best to go by...



If your tourney finishes can be exported to excel it wouldn't be difficult to write a sheet that would sort that data into a similar structure.

 

[darthdimsky]

Thanks K. I've posted this question in the PT4 support thread to see if there's a way to arrive at that using the software itself.

But putting it up in a spread and grouping them on %s using a simple IF THEN ELSE seems to be most reliable way of going about it.

PS - Just edited the OP with a link to the description for Net Expected BB. It's a really awesome measure.

 

[terryk]

Damnnnn,,,Dims! Nice stats! :icon_salu

 

[darthdimsky]

Last year yea. This year no so much LOL.

 

[darthdimsky]

Was able to export a small sample and work on it in Google Spreadsheets. Figured out the vlookup for this bit of work.

 

[DaveE]

Looks good :top:

 

[BetterNot]

Nicely done! I also usually look at the percentages it gives a clearer picture than only your place

 

[darthdimsky]

Yep. Also found out another major ttracking site (officialpokerranking.com) also uses the same metrics.

Suprised by the few replies. Thought there'd be a few other MTT grinders out there who might contribute. Maybe I should consider posting it in the "Learning Poker" section?

 

[AlexTheOwl]

Net Expected BB won: This measure is the best I've found so far in PT4. It isn't result oriented. It measures the expected outcome of your hands. While comparing against actual BBs won it clearly takes into account variance and is probably the best indicator of how you're performing. Ref first screenshot. Next Expected outcome is explained in more detail with an example here.

This a quote from the moderator of the Poker Tracker support forum that you linked. It is in response to a question about how this stat is calculated.:

"The short answer is that when you or your opponent is all in, your equity is multiplied by the pot size, and that is the amount you are expected to win in the long run. Please note that this will almost never be the amount that you actually in fact win."

So this stat only seems to measure how well you would perform in all-in situations if there was no variance. That's an interesting stat to know, but it is not a general measure of how well you are performing in tournaments.

 

[darthdimsky]

This a quote from the moderator of the Poker Tracker support forum that you linked. It is in response to a question about how this stat is calculated.:

"The short answer is that when you or your opponent is all in, your equity is multiplied by the pot size, and that is the amount you are expected to win in the long run. Please note that this will almost never be the amount that you actually in fact win."

So this stat only seems to measure how well you would perform in all-in situations if there was no variance. That's an interesting stat to know, but it is not a general measure of how well you are performing in tournaments.

True. It's not a true indicator of how well you're doing in MTTs. Though it does indicate whether you're making the right decisions overall. It is a good measure to determine whether you're truly going through a downswing or whether you're just playing badly.

 

[AlexTheOwl]

True. It's not a true indicator of how well you're doing in MTTs. Though it does indicate whether you're making the right decisions overall. It is a good measure to determine whether you're truly going through a downswing or whether you're just playing badly.

I agree that this stat can be an indicator of a downswing.

It is an indication of whether you are making the right decisions in hands that end in all-ins. The great majority of hands do not end in all-ins, and it tells us nothing about those hands. Although of course all-in hands tend to be the most important ones.

Also, have you considered that this stat can be either too low or too high?

If the stat is too low, then you are getting it in with the worst of it too often, which is bad for obvious reasons.

If this stat is too high, it means you should be winning a very high percentage of your all in all-in situations. That sounds like a good thing, but:

If that is the case you are probably only getting it in when you are fairly certain that you have the best hand.

But, as I am sure you know, sometimes you should be all in even when you have serious doubt that you have the best hand, such as:

- going all in because pot odds justify it
- going all in as a bluff or semi-bluff that will succeed often enough to be profitable. When all in bluffs work they will not affect this stat, but when they fail they will bring this statistic lower.
- going all in when you don't have the nuts or close to the nuts, but are still ahead of your opponent's range.

A player who never goes all-in when behind no matter what the pot odds, never bluffs all-in, and never goes all in without the nuts would have a very high value for this stat, but that player would have serious leaks!

I don't know what values would be good for this stat. Do you have a target value in mind for it?

 

[darthdimsky]

I agree that this stat can be an indicator of a downswing.

It is an indication of whether you are making the right decisions in hands that end in all-ins. The great majority of hands do not end in all-ins, and it tells us nothing about those hands. Although of course all-in hands tend to be the most important ones.

Also, have you considered that this stat can be either too low or too high?

If the stat is too low, then you are getting it in with the worst of it too often, which is bad for obvious reasons.

If this stat is too high, it means you should be winning a very high percentage of your all in all-in situations. That sounds like a good thing, but:

If that is the case you are probably only getting it in when you are fairly certain that you have the best hand.

But, as I am sure you know, sometimes you should be all in even when you have serious doubt that you have the best hand, such as:

- going all in because pot odds justify it
- going all in as a bluff or semi-bluff that will succeed often enough to be profitable. When all in bluffs work they will not affect this stat, but when they fail they will bring this statistic lower.
- going all in when you don't have the nuts or close to the nuts, but are still ahead of your opponent's range.

A player who never goes all-in when behind no matter what the pot odds, never bluffs all-in, and never goes all in without the nuts would have a very high value for this stat, but that player would have serious leaks!

I don't know what values would be good for this stat. Do you have a target value in mind for it?

Dug around a bit more. It does appear to be irrelevant, like you suggested in your first post, for the most part. PT4 mods have explained it a bit more for users here. Might be useful to understand how I fare in different stages of a tourney by filtering it on blind levels, based on suggestion here.

At one point I mistakenly believed this applied for all hands because clicking on the line actually pulled in hands that were folded pre even.

- going all in as a bluff or semi-bluff that will succeed often enough to be profitable. When all in bluffs work they will not affect this stat, but when they fail they will bring this statistic lower.
- going all in when you don't have the nuts or close to the nuts, but are still ahead of your opponent's range.

In both instances, if I've understood it right, the next expected BB values should rise because the hero's ahead in both instances and technically gets a bad beat after shove.

 

[iwont20]

winrates by stacksize and by positions

 

[darthdimsky]

winrates by stacksize and by positions

I've barely used the reporting functions in PT4. But with the reading and the suggestions I'm getting more and more used to it. Didn't even know you could report on stack sizes till this thread.

I've created two very basic reports around PF Stack sizes and positions. Will upload screenshots as soon as I can (facing an error uploading them atm).

If anything, the suggestions and reading has helped get a better grasp of reporting and analyzing for leaks.

 

[AlexTheOwl]

- going all in as a bluff or semi-bluff that will succeed often enough to be profitable. When all in bluffs work they will not affect this stat, but when they fail they will bring this statistic lower.
- going all in when you don't have the nuts or close to the nuts, but are still ahead of your opponent's range.

In both instances, if I've understood it right, the next expected BB values should rise because the hero's ahead in both instances and technically gets a bad beat after shove.

I apologize, you are correct about going all when you don't have the nuts, since the denominator is hands played. If the denominator was hands where all ins are called, then I would have been correct (and the stat would be more useful, though still in a very limited way).

Here is my thinking on bluffs. I could be wrong:

When the bluff is successful, the villain folds. The villain's cards are not exposed (maybe this varies depending on the poker site you play on?). Since the villain's cards are unknown, PT4 cannot do an equity calculation, so this hand is not included in net expected BB values.

When the bluff fails, the villain calls. Since this is a bluff, by definition, you are behind most of the time. Net expected BB is lowered.

So successful all-in bluffs are not captured in this stat, while failed ones are included.

PT4 is a great product, but this is a deceptive and almost useless stat. It should at least be re-named to reflect that it applies only to all-ins where two or more hands are exposed.

ROI and ITM have the problems that you pointed out in your original post, but they are still the best measures we have of overall performance in tournaments.

 

[darthdimsky]

ROI and ITM have the problems that you pointed out in your original post, but they are still the best measures we have of overall performance in tournaments.

The ROI and ITM combined could be the best measure.

• High ITM & High ROI - MTT player crushing whatever stakes he's playing
• High ITM & Low ROI - slightly winning MTT player with more focus on min cashing
• Low ITM & High ROI - possibly a winning MTT player with more focus on deep runs
• Low ITM & Low ROI - more likely a losing MTT player

I'm sure there is a way to arrive at a metric combining both of these numbers. And if used with those grouped positions will be very powerful.

 

[AlexTheOwl]

The ROI and ITM combined could be the best measure.

• High ITM & High ROI - MTT player crushing whatever stakes he's playing
• High ITM & Low ROI - slightly winning MTT player with more focus on min cashing
• Low ITM & High ROI - possibly a winning MTT player with more focus on deep runs
• Low ITM & Low ROI - more likely a losing MTT player

I'm sure there is a way to arrive at a metric combining both of these numbers. And if used with those grouped positions will be very powerful.

Yes! If you take on this challenge, I will be very interested to see what you come up with.

 

[darthdimsky]

Yes! If you take on this challenge, I will be very interested to see what you come up with.

I honestly doubt I'll be able to make anything meaningful unless I had access to a large db of player with the relevant numbers. That said it's probably better in the hands of more mathematically inclined players like Vinnie (you should check his posts out!)