Practicing Ranges

Some folks have been asking about how to practice putting villains on ranges. I admit, I have sucked at this for a long time, but I also have the advantage of knowing a bit of learning theory. I've tried several things to help organize my range estimates at the tables. This is the process I'm currently using.

Chunking

Chunking is like phone numbers, grouping several items into groups of 3 - 5 so the brain can process / remember them more easily. I created 9 chunks organized according to strength and postflop playability. You might argue with it - that's fine, make your own. The key features of these chunks are:

1. Each chunk is approximately 5% of all NLH starting combos.
2. The hands within each chunk tend to get played in similar situations (like the Axs / sc's get 3bet light a lot, for example).
3. The chunks are arranged top-to-bottom according to strength (roughly).

Here are my chunks, along with the names I identify them with at the tables:

Premium: AA, KK, QQ, AK, AQ, KQ
BW-J: AJ, AT, KJ, QJ
pp's: JJ - 22
BW-T: A9, QT, KT, JT
Axs/scs: A8s-A2s, T9s-65s, J9s-86s
Axo: A8o-A2o
Kxs/Qxs: K9s-K2s, Q9s-Q2s
9's: K9o, Q9o, J9o, T9o, 98o, T8o

If you want different groups, you can rearrange this to your heart's content. Now that we have them, wtf do we do with them?

Ranges at the Table

Here's a simple example. I get AQ in the CO, open for my standard raise and get flatted out of the blinds by a TAGG wannabe, say, 18/14/2.5 with 3bet ~ 2.5% . All right, I'm trying to estimate his range. That 2.5% 3bet means half the "premium" is gone. I don't rule out AA or AK all the way - I think of it as about half the total combos gone since, at my stakes, some folks will flat KK+ (that didn't happen much at 10nl, btw).

Now, if I'm in a tight spot, I can check popup stats to get exact flatting percentages. Here, what I do to estimate the percent of hands he'll flat is take the "gap" in VP$P and PFR (18 - 14 = 4) and multiply it by 2.5: 10% estimated flatting range. For TAGG's who are a bit positionally aware, I use "2.5x gap" for Blinds/EP and "3x gap" for CO/BTN. It gives a decent estimate of the numbers of hands I need to include, for reasons we can discuss later in the thread, if anyone cares. For the moment, let's assume he'll call with the 10% just below his premium 3betting range.

OK, so I need 10% of hands not including the "half premium" chunk he'll 3bet. So I grab all the pp's chunk, obv, plus the "worst" half of the premium chunk and the best half of the "Broadways w/ J's" or BW-J chunk. That's about right. Just for clarity, I'm thinking that AQ, KQ, AJ, KJ and JJ - 22 are in this flatting range.

I also include a bit of "junk" or wiggle room in the range, just to admit that he might do something different or simply lose his mind and flat with 65s for reasons known only to him. That's why I don't do much with sooted cards in the top 4 chunks - I need simple and easy-to-use chunks and can toss in a couple sooted combos when needed, usually keeping track of them with the "junk" part of the range.

When the flop comes T83 rainbow, I'm cbetting pretty much every time (regardless of what I actually have) since these cards only hit the sets and JJ which are a pretty tiny portion of his range. I wanted y'all to see the end result of the idea: the ability to grab and deal with chunks or half-chunks instead of individual cards.

Practicing Ranges Estimation

Step 1 is to really understand your HUD stats and what they mean. They give us usable figures so we know how many combos to throw into the preflop range. (Remember, we always add few "extra" combos or "junk" so we don't get crossed up when the action fails to conform to our estimate.) That's a whole other thread, so I won't mention much about HUD stats here.

Step 2 is to take HH's and work through them, pausing on each street to formalize the range. After the flop, I'm throwing combos out of each chunk, and trying to keep track of the parts. It's not easy and does take time to master when I have to do it in my head. That's why I practice away from the tables.

Learning Theory

Each of us has various learning styles, usually one we prefer and one or two others that we can manage OK. I'm a visual-spatial learner, and most people enjoy kinesthetic or tactile (i.e. "hands on") learning. Basically, the more senses we get involved, the more we're likely to retain. So I made this chart:

Poker Ranges

You can see each box has a chunk in it, so each is about 5% of combos. I take about 20 pennies, and I start off preflop marking whole chunks by placing a penny in the top corner of a chunk. If it's a half chunk, the penny goes in the bottom corner. Postflop, I tend to place pennies over the combos inside the chunks that I'm eliminating, so I see only the combos that are left. I keep going street by street, moving the pennies around and narrowing the range further.

The process is visual, kinesthetic, easy-to-use and workable in terms of usefulness translating to actual game play. OK, so much for learning theory.

Finding HH's to Practice With

The final suggestion I'll make is to use PT3 or HEM to "profile" someone you have 500+ HH's on. I find someone with "typical" stats, something like 24/16/3 that I see a lot of at my tables. I filter using the following criteria: "Showdown = TRUE" and "Final Pot > 40bb." My rationale is that these hands will play out over multiple streets, so I'll have several chances to narrow the range.

I can often find villains with 1k+ HH's, and this filtering generally yields about 10 - 12 hands. Usually, I played postflop in less than 2 or 3 of them (some of which I remember, so I don't want to use them unless I'm analyzing my own play).

I start the replayer and play the hand like I'm this Hero. I estimate the ranges for Hero with pennies and paper, and estimate the ranges for whatever villains join in based on their HUD stats (but do it my head). I grade my session. Out of 10 HH's, I think I ought to be down to 5 - 6 combos (and be right!!) about half the time. Sometimes, you just can't narrow it down too much because someone just shoves a bluff. But even there, I mark an effort as FAIL if I can't guess within 10 combos. If I can nail down as many as 2 HH's to 3 combos or less, that's A work.

Each HH I analyze takes 3 - 5 minutes, depending on how wide the preflop ranges are and how much information I can get from the action, board texture and other reads. So I can do 8 - 10 in half an hour. I try to do one range practice session per week.

Example
Here's an easy one to get you started from my last practice session. The "Hero" I'm analyzing in this hand is the BB and has these stats: 22/16/3 over 900 hands with 11% 3bet from BB over 64 chances. Her "folds to cbet" is 51% and "raises cbet" is 11%. Her "folds BB to steal" is 77%.

The villain in the hand is the SB who is 40/10/1.3 over 90 hands with "steals from SB" of 30%.


$0.25/$0.5 No Limit Holdem
5 players
Converted at weaktight.com

Stacks:
UTG ($120.85)
Hero (CO) ($90.80)
BTN ($64.45)
SB ($72.75)
BB ($98.25)

Pre-flop: ($0.75, 5 players) Hero is CO
1 fold, Hero folds, 1 fold, SB raises to $2, BB calls $1.50

Flop: ($4, 2 players)
SB bets $4, BB calls $4

Turn: ($12, 2 players)
SB bets $12, BB calls $12

River: ($36, 2 players)
SB bets $20, BB calls $20

Final Pot: $76



If you're interested, work through the HH using my chunking method or any other organizational system that helps you. Discuss each street, and narrow the BB's range to less than 5 combos. Bonus points if you can narrow it to 3 or fewer combos for the BB.

Focus on BB initially, and we'll discuss the SB's range later.

interesting idea, i like it.

Anyone who hasn't done their own range work should go through Robb's approach or sit down and figure out an approach of their own.
If you haven't studied ranges already then this is way more long term +++EV than another one hour session grinding $5nl...

when i first started to trying out what the various percentages meant, the first step was to know % for any pocket pair, any suited combo, any non-suited combo and any combo irrespective of suitedness. The 4 corner-stones of RANGes 101.

I still play around with this, some sample blocks (chunks in Robb-parlance )

AK/QQ+
77-JJ
22-66
broadways
SCs
SGs
Axs
etc, depends what i'm doing it for.

preflop we have 11% 3b
let's say QQ+,AK is the value range
Ax (suited or otherwise), 22-66 are the bluffing ranges
but that's 11.9% + 2.6% value range so we'll just say he folds his bluffing range 1/4 of the time so as not to get owned (and the hands we fold are Axo almost every time - the bottom of our range)

now we need 12% calling range
JJ-77,AQs-ATs,KTs+,QTs+,J9s+,T8s+,98s,87s,76s,65s,AQo-ATo,KJo+
possible QJ, but I thought in the mindset of this player it's a real stretch

continuing range on the flop:
TT,AcTc,AdTd,AsTs,KcQc,KdQd,KsQs,KTs,QcJc,QdJd,QsJ s,QcTc,QdTd,QsTs,JcTc,JdTd,JsTs,AcTd,AcTh,AcTs,AdT c,AdTh,AdTs,AsTc,AsTd,AsTh,KcQd,KcQh,KcQs,KdQc,KdQ h,KdQs,KsQc,KsQd,KsQh
I've taken out some raising hands and QT of hearts probably slowplays this because of nut hand + nut draw

by the turn a calling hand has to be pretty strong, and QT raises to protect against at least ONE of the flush draws
AcTc,AdTd,AsTs,KdQd,KdTd,QdJd,JdTd,AcTd,AcTh,AcTs, AdTc,AdTh,AdTs,AsTc,AsTd,AsTh

river confuses me, if he made his flush he should make a thin value raise with a flush
so I can only see him holding AT

there are other hands, sure, but I am going not just for BONUS MODE
but I am going for NIGHTMARE MODE: one hand only

Originally Posted by iopq

now we need 12% calling range
JJ-77,AQs-ATs,KTs+,QTs+,J9s+,T8s+,98s,87s,76s,65s,AQo-ATo,KJo+
possible QJ, but I thought in the mindset of this player it's a real stretch

I like pp's 22 - 66 in flatting range. I could make a case for QJ, but see your point.

it's a style issue, we don't know how BB plays
given his opponent I'd have to agree with you because we probably have little fold equity with a 3b and flatting becomes better
fortunately, that doesn't change the rest of the hand
unless we're going to go out there and say well he's going to flat Axs for the same reason in which case that totally rapes my hand ranges

Thanks for the post, good stuff as usual Robb I haven't taken the in-depth practical approach that you do.
I used to simply practice by continually calling out hands (out loud) of the folks involved in the hands and modifying/narrowing them as I went. Ofcourse this doesn't work so good after about 4 tables.
I've created a few starting hand chart with some % ranges color-coded in a color hierarchy (like any starting hand chart you normally see) with 4%, 8%, 12%, 20% - you could go higher of course, by I find beyond that the range isn't narrow enough to knock very much out. But anyways, while looking at my HUD, if I see a guy is pfr 12% and he makes a standard raise from mp, I can just glance at the 12% category and have a pretty good feel for which hands are potentially there.
I certainly don't use this all the time, but it's good practice and helpful to your overall game.

Robb, can I make a suggestion for exercises of this sort , that you title the thread "Homework: ranges " and try and get spenda or stax to lock the thread and temporarily sticky it,once you put the "homework " question up and give a couple of days for us noobs to do the analysis themselves and possibly pm the answers to you. You can then post the replies and attribute them to the senders once the deadline has passed.
Everyone then has to work through the theory themselves rather than cribbing ideas from other peoples answers. Then everyones answers could be critiqued to see where they went wrong , what they did right etc.

I had a long talk with a huge fish about this HH last night, I think knowing one of the two hands is huge.

Originally Posted by bigspenda73

I had a long talk with a huge fish about this HH last night, I think knowing one of the two hands is huge.

For this exercise I will ignore that 9s and 7h are dead to simplify the process.

I'll give this a try. I'm going to go out on a limb and make some pretty critical statements, but I think you'll read them in the spirit it's intended.

First off I'll be perfectly clear and state that range practice is to me clearly the biggest thing I could/should do to improve my game - and I'm not doing it. Also don't have much time to play though.

Before diving into the range practice proper I'd like to dwell a bit on the players in the hand and some other things. BB is an interesting profile in that he continues to steals 23% of the time and has a vpip of 22%. This almost suggests someone who is not positionally aware - or perhaps more accurately someone who might not be playing a positionally aware range, but who might still play his range in a positionally aware way! It's possible of course that he plays 10% UTG and 35% on the BTN and that his BB range happens to be the same as his average vpip. Anyway, he plays 23% of hands - which 23%? I can go with premiums and pp's, but I'm not sold on having BW-J and BW-T so high on the relative hand strength order.

I'll divert here and discuss briefly a few things from BobboFitos DVD set that I got from PokerStars not too long ago. When I saw his tiering of starting hands I similarly did not fully agree/support and I present a rough approximation of his as a counterargument to yours. His T1 is pocket pairs, AK and Axs - T2 is suited connectors to 65s and suited gappers down to 75s and AQ. T3 is the rest of broadways (suited or no), the lower suited connectors/one gappers, offsuit connectors and Axo. The principle behind is Axs in T1 and all of T2 over T3 is hands with big hand potential (straights/flushes) being more playable than hands with high card value/showdown equity/hot and cold equity.

This aside meant to get you thinking - should you always consider your chunks in the order listed here?

Ok, back to the hand in question, I need to pinpoint 23% that my BB is playing against SB. SB is likely on a steal (two steals per proper hand) and BB probably knows this. 23% is about 310 hand combinations. I'll pick hands with big hand potential first. All pocket pairs (78), All suited aces (48), AK (12), AQ (12), KQs-65s (32), KJs-75s (28), KTs-J8s (12), AJo/ATo/KQo/KJo/KTo/QJo/QTo/JTo (96) - 318 hand combinations.

We know the BB doesn't widen his hand range in blind defense, but he does raise quite vigorously in blind defense. Almost half the time. The question again is, which hands does he choose to 3bet rather than call? If he was following randomization by hand equity, ABCD theorem or the golden ratio (same thing btw) he'd probably be raising AA-QQ (18), AK(16), T9s-65s (20), T8s-75s (16), Q9s-J8s (8), KJo (12), KTo (12), QJo (12), QTo (12), JTo (12), for 138 hand combos.

I'm not sure what other logical way someone can have of putting together a 3betting range. I guess we could put JJ/TT and AQ in there some of the time as value with less bluffing hands, but I don't think it's really that far off.

This makes my flat calling range for BB:
JJ-22 (60), AQs-A2s (44), AQo-ATo (36), KQs-JTs (12), KJs-J9s (12), KQo (12), KTs (4) - 180 hand combinations

The board is scary for any hand and calling a pot sized bet seems like a really odd choice. You almost have to have both equity and outs to call. I could see even weak top pairs folding here (like A7s). I don't see TT-22 calling. While BB's fold to cbet is 51%, I'd put the chance of folding to a cbet on exactly this flop maybe a bit higher than the norm. It's a dangerous flop where BB will only continue with a hand. I don't expect much light floating on this flop.

Potential raising hands: (nuts) QTs, JJ, QhJh, JhTh (bluffs) TT
This happens to be 15 hand combinations which isn't too far removed from 11%. Needs be said that if either of these hands decided to slowplay by calling the flop they would raise the turn for sure, so I'm just counting them out at this time.

My flat calling range:
AQ-AT, KQ, KJs, KTs, QJs, JTs

Turn is unlikely to improve anyone directly but it does put out another flush draw. From this point basically any diamond, any heart, any queen, any T and any card that pairs the board is scary as hell. SB is threatening another bet on the river and BB can only call if he is showdown bound. On the flop I included a lot of hands that had pair plus straight draw, but on the turn most of those hands have to fold. I think we're pretty much down to AQ-AT, KJs.

River pairs the A and completes the flush. AJ would raise, KJs would fold - I can see AQ and AT.

Right now I have trouble putting SB on anything other than KJ, but I haven't really thought through SBs situation.

Edit note: Forgot a lot of heart flush draws in my flop range. I'm guessing they're folding the turn also though.

Upon reflection, I won't rule out KdQd for BB. I had it in my flop calling range (as all KQ) and with the flush draw I guess he might be tempted to call turn as well as any diamond (not ace), ten or king improves his hand.

For SB, I initially didn't like the river bet with a full house. Upon reflection I can't rule out full houses or quads because they want value (to be called) and too strong a bet can elicit folds. So with full houses in SBs range I expand that with AK, AJ, AA, KK, JJ. And since my basic thought for flop and turn was that he's betting with two pair or better, I have to also include QT - which is also consistent with the river bet. With the bet size I think he has a lot more value than bluff/semi-bluff hands in his range, but if we were to include some semi-bluff I guess they would be the AQ/AT variety or perhaps lower heart flush draws. Lower heart flush draws don't make this river bet and AQ/AT might.

Modified BB range: AQ/AT/KdQd
Modified SB range: AA/KK/JJ/AK/AJ/KJ/QT/AQ/AT

the problem with KdQd that makes it discounted is it's so tempting to raise the river since SB is repping trips

Originally Posted by Robb

Premium: AA, KK, QQ, AK, AQ, KQ
BW-J: AJ, AT, KJ, QJ
pp's: JJ - 22
BW-T: A9, QT, KT, JT
Axs/scs: A8s-A2s, T9s-65s, J9s-86s
Axo: A8o-A2o
Kxs/Qxs: K9s-K2s, Q9s-Q2s
9's: K9o, Q9o, J9o, T9o, 98o, T8o

AA & KK are so absurdly good that you can't chunk them with any other hand.
QQ is still really strong
JJ is under-rated
AK and TT are of about comperable strength, you just get delt AK more often.
99-66 are much better hands than 55-22.
55-22 lose a lot of value against people know when to give a lot of post-flop action. They don't crush boards the way the bigger pairs do.

Axs and suited connecters are VERY VERY different hands in the unlimited hold'them.
Axo is dogshit with any reasonable sum behind and utter crap without position.

Axs
o Play very well mulit-way with lots of money behind. You will often trap people second best when you nail a flop hard.
o Very position sensitive, partilarly when you flop a flush draw or hit just your Ace.
o Don't often have hidden strength when you rep stronger hands with them.

Suited connectors
o In a multi-way pot (5+ players) you will make a lot of second best hands.
o Work really well in 2-3 way pots against players who know how to fold when you mix them in with stronger hands. Hands that would have been second best become the nuts when a wider range gives action and you kick hands like k7s out of the pot.
o Aggression with initiative isn't as position sensitive.

Thank you thank you thank you.

My list of things to do include something like this - for each playable starting hand outline for myself the things I need to look for to determine if it is profitable to play it.

An addendum I'd make to the Axs hands is probably that AKs-ATs > A5s-A2s > A9s-A6s.

One thing I didn't really get into is also that when we ask ourselves which hands are in a range we tend to do it the way Robb described here and Renton did in his ABCD theorem post: By imagining for the sake of the exercise (to make it manageable) that there is a single true way to rank hands from strongest to weakest and then we can take the very top as our nut range (monsters), next section as our showdown range and the next section as our semi-bluff range and the last section as our folding range. This is simplistic and flawed because it establishes habits such as thinking that AJo is always a better hand than 76o where situationally this may not be true.

Fnord makes a terrific point in saying how strong certain drawing hands are based on how multiway a flop is, but similar principles must logically apply for villain's playing tendencies. If he chooses to play certain types of hands very aggressively for instance - like a player who overplays overpairs and always stacks off with them regardless what's on the flop: 76o might be better against him than AJo because with 76o if you hit two pair or your straight you know you're good, where as with AJo you can't know if you've been outkicked or is up against a set.

So basically when I do a hand range practice for a given villain I should consider his playing tendencies and determine based on that whether he is more likely to have broadways over suited connectors or naked unsuited aces over other cards etc - then I need to consider which hands he will fold or play for small pots and which hands he'll stack off with - then I need to consider which starting hands I can play that is likely to make which type of showdown hands (if they hit) and decide to play the starting hands that are most profitable against the playing tendencies of this particular opponent, and based on his tendencies create an EV hand ranking against this particular opponent. Then I promote the top of this range to my nut range (in preflop terms this means 3bet with intent to stack off for value), decide on the profitability cutoff point, weigh the factors that decide how big a ratio of bluff to nuts I need in my 3bet range and add in that amount of weak EV hands to my 3bet range and flat call the rest.

I think that is the theoretically correct way to do it, and I'm of course pretty incapable of doing it because I haven't done anywhere near enough range practice. Also something that's on my todo list is creating a library of tendencies. My hope there is to familiarise myself more with what tendencies are possible and find an easier time remembering them as applicable for a specific player when I have them mentally steeped in the context of how they are exploitable and what range/play adjustments are appropriate for this specific tendencies. This should make note shorthand more effective - and in turn make it easier/more possible to make a proper detailed range practice.

Actually I'd say AKs > JJ > AKo > TT
A7s > A5s > A6s

Originally Posted by iopq

Actually I'd say AKs > JJ > AKo > TT
A7s > A5s > A6s

WTF cares? You've totally missed the entire point of this thread with your retarded savant's focus on irrelevancies.

Originally Posted by Robb

Originally Posted by iopq

Actually I'd say AKs > JJ > AKo > TT
A7s > A5s > A6s

WTF cares? You've totally missed the entire point of this thread with your retarded savant's focus on irrelevancies.

This coming from you? LOL