Comparing Simple Strategies

Poker Strategy
182

Wow… pulled it out of the fire again, although only barely. Pretty exciting too. It came down to the final 120 hand session, going in about $200,000 down still. I got stacked with about 50 hands to play (making a really ambitious call of an all in, slight over-bet of the pot that would have only beat a bluff), and it felt like all hope was lost. With about 10 hands to go, on the same table, I doubled up (I’d reloaded after getting stacked) breaking even on that table, and small winnings on the other two tables were enough to push me into positive territory.

This run was Balance V, for high Value. I basically cut the bluff targets in half from the prior run, to leave it quite value heavy. Again, staying close to the target ratios was not easy, as it is not like you can force a correcting play on the next hand or two when you notice that you are skewing one way or the other, and so I did miss some of the desired targets again. Still, I definitely achieved the main objective of increasing the ratio of value to bluffs relative to the prior run.

  • Flop
    • pot sized bet target: 2 value to 1 bluff
    • pot sized result: 22 value to 9 bluffs
    • 1/2 pot bet target: 4 value to 3 bluffs
    • 1/2 pot results: 25 value to 18 bluffs
  • Turn
    • pot sized bet target: 3 value to 1 bluff
    • pot sized result: 13 value to 6 bluffs
    • 1/2 pot target: 2 value to 1 bluff
    • 1/2 pot result: 19 value to 9 bluffs
  • River
    • pot sized bet targets: 4 value to 1 bluff
    • pot sized result: 10 value to 2 bluffs
    • 1/2 pot target: 3 value to 1 bluff
    • 1/2 pot result: 18 value to 5 bluffs

I often get surprised by the results of these runs. I kind of thought this had a fair chance of outperforming the first Balance run, as there would still be enough bluffs that I’d show down a few, and hopefully that would be enough to get a lot more calls for my value bets. What ended up happening: especially in the first half, I got value owned a lot, making value bets into what turned out to be better hands. That will always happen from time to time, and of course doesn’t mean the bets were not value bets, or that I shouldn’t have made them… it’s just another way you can and will run bad sometimes. On the flip side, I obviously got lucky a time or two also.

But I have felt in general like my bluffs have been more profitable on their own than I’d expected going in. It might just be that a lot of players see my rank, and think something like, “gee, I’m just going to fold unless I have a monster.” So it might be that people are folding more against me than against most of the other players on the level… I’m not really sure. Also, in general, if you apply a lot of pressure, you’ll often get rewarded with your opponents making bigger mistakes as you take them out of their comfort zone. But as the name of these latest strategies suggest, that pressure usually needs to be kept somewhat balanced.

Balance V results:
Tables played: 1k/2k NL Holdem 9 max
Hands played: 1,000
Chips won: $219,038
BB/100 hands: 10.95

183

OK… kind of an embarrassing stretch here… I started 2 prior runs at 2k/4k and aborted them, as I just didn’t seem to have the mental focus to stick with the strategies I was intending to test out. Now, most of the way through a 3rd run of what was supposed to be Laggy TAG, I’ve kind of run into the same problem, and haven’t really consistently followed any of the rules for that strategy, except perhaps the pre-flop opening range (and haven’t really stuck completely to that, even).

So this has really turned into a Night Rain run, where I’m doing whatever I want. On the whole it feels like I’m playing unusually poorly, but getting lucky enough to more than offset that. Given my general lack of focus for the last few days, I think I’ll end up calling this Light Rain. So the run has mostly been whatever I’ve felt like doing at the moment, but below are a few areas where my play could be defined at least part of the time:

  • as I start tables and find myself heads up, I’ve generally played very passively (though there were some exceptions to this also), as I didn’t want to chase people off
  • in the later parts of the run, I’ve started deviating more often from the opening ranges described in Laggy TAG, generally making each of the sections (100%, 50%, 25% and 10%) a bit broader
  • I used the wacky c-betting algorithm briefly (it seems silly, but I’ve actually found it surprisingly effective in the past), but maybe only for the first 200 or 300 hands

This all raises an interesting question: what are some things you can do when you feel like your not mentally able to play your normal game, for whatever reason?

  • take a break
  • drop down in stakes
  • try some new things
1 Like
184

All three.

185

Play without the hassle of collecting data for awhile. Poker is a game… just play for fun before you burn out.

Meanwhile, figure out where and when and why each strategy worked, and where, when, and why it didn’t. Extract from this some general guidelines tailored to the stakes you play and the kinds of players you usually face.

The player pool here isn’t playing optimally, (dirty little secret, no player pool is) so figure out what will work here, where you play! Just have fun with it!!!

2 Likes
186

Light Rain results:
Tables played: 2k/4k NL Holdem 9 max and 6 max
Hands played: 1,029
Chips won: $583,215
BB/100 hands: 14.17

So this was really mostly just me playing my normal game, and being pretty close to having a losing session over 1,000 hands. I also called an all in raise over my bet on the turn with a non-nut hand near the very end of the session. If my opponent had happened to have an even better hand, or if I’d gotten unlucky and they’d hit one of two outs they had, that would have been enough to make this a losing session.

So my streak of winning 1,000+ hand sessions has piled up to 29 in a row, but it’s taken a lot of luck to get that many, and even if I was playing my best poker rather than mostly simple strategies, I’d bet against myself being able to repeat that. I’d also note that in general, since I started playing on this site, while it is not unusual for me to have several days in a row of net winning results, getting a week without a single losing day is always a bit special, and I don’t think I’ve ever had a full month without a single losing day.

Results summary:

  • Pre-Flop Hammer @ 1/2: 222 BB/100
  • Super Nit @ 1/2: 856 BB/100
  • Passive Fish @ 2/4: 73 BB/100
  • Lazy Limper @ 2/4: 372 BB/100
  • Value Village @ 5/10: 220 BB/100
  • Pre Flop Hammer @ 5/10: 241 BB/100
  • Value Village @ 10/20: 176 BB/100
  • Lazy Limper @ 10/20: 244 BB/100
  • Pressure Cooker @ 25/50: 139 BB/100
  • GAG @ 25/50: 300 BB/100
  • Robo TAG @ 50/100: 131 BB/100
  • PFP @ 50/100: 130 BB/100 (guest run)
  • GAG Cook @ 50/100: 190 BB/100
  • Pre Flop Hammer @ 100/200: 2 BB/100
  • Get da Fishy @ 100/200: 104 BB/100
  • LAG Problems @ 100/200: 125 BB/100
  • PFP @ 200/400: 30 BB/100
  • Orphaned Pot Adopter @ 200/400: 49 BB/100
  • Value Village @ 500/1,000: 109 BB/100
  • Balance @ 500/1,000: 153 BB/100
  • Balance V W 1k/2k: 11 BB/100
  • The Maniac @ 1k/2k: 39 BB/100
  • Light Rain @ 2k/4k: 14 BB/100
  • Random Fool @ 2k/4k: 26 BB/100
  • Multiple Personality Disorder @ 5k/10k: 66 BB/100
  • Small Ball @ 10k/20k: 3 BB/100
  • Super Nit @ 20k/40k: 58 BB/100
  • Laggy TAG @ 50k/100k: 27 BB/100
  • Night Rain @ 50k/100k: 77 BB/100
187

OK, the latest run has been going at 5k/10k, and I’m just a little bit over half way through it. I’ve been off to a good start, but this will be a high variance strategy, and I’m also about to handicap myself by revealing the mechanics.

This is Balance B, for high bluff balance. I can take any lines I want, but I’m striving for certain targets between value and bluffs for each bet size on each street. In prior Balance runs, I’ve only used 1/2 pot and pot sized bets. Here, I’m using three bet sizes:

  • s: small bets usually between 20% and 40% of pot
  • n: normal bets from 50% to 100% of pot, mostly clustered in the 60% to 80% range
  • o: over-bets, typically in the 110% to 300% range

Flop Betting:
Targeting 1 value to 3 bluffs for small bets, 1 value to 2 bluffs for normal bets, and 2 value to 3 bluffs for over-bets. Assuming I’m c-betting, I’ll typically use the smaller size with 100% of my range on dry flops, and will make normal sized bets with part of my range into wetter flops. Over-bets will be fired exclusively into wet flops with either premium made hands or strong draws.

Turn Betting:
Aiming for 1 value bet to 2 bluffs with small bets, 2 value bets to 3 bluffs with normal sized bets, and 1 to 1 for over-bets. Here I don’t really have a super simple way of defining my sizing, but the smaller bets will again mostly be into dry-ish boards, or with more opponents, while the larger bets will typically reflect either made hands or good equity.

River Betting:
2 value to 3 bluffs with small bets, 1 to 1 for normal sized, and 3 value to 2 bluffs for overbets. I’ll mostly be picking sizes to fill the holes as my ratios get out of wack, which is not a strategy I’d typically recommend, lol. So it will kind of be like, “ok, it looks like I don’t have enough normal sized bluffs… ok, here goes.” If I have time to think about it, when I’m bluffing, I might ask myself, “what part of my opponents range am I hoping to get folds from.” When value betting, I might similarly consider what weaker hands I’m most wanting calls from, and what bet sizing is likely to work best for that.

Edit on Flop C-betting:
Small sized c-bets into dry boards with 100% of range assumes 1 opponent, and even then I’ll pick some spots to check, so c-betting frequency will probably be closer to 85% to 90% with one opponent, and will drop further as there are more opponents (though I’ll still be firing plenty often).

1 Like
188

Oh, I should also mention… this one is loads of fun to play, a bit like a roller coaster. For those of you that are gamblers by nature, the high density of big hands makes for an exciting ride.

189

OK, results are in.

Balance B results:
Tables played: 5k/10k NL Holdem 9 max and 6 max
Hands played: 1,001
Chips won: $10,957,850
BB/100 hands: 109.47

Final frequencies:
Flop (s1/3, n1/2, o2/3 targets):

  • small bets: 8 value bets to 27 bluffs
  • normal bets: 17 value bets to 28 bluffs
  • over bets: 5 value bets to 6 bluffs

Turn (s1/2, n2/3, o1/1)

  • small: 4 value to 9 bluffs
  • normal: 10 value to 16 bluffs
  • over: 11 value to 8 bluffs

River (s2/3, n1/1, o3/2)

  • small: 7 value to 8 bluffs
  • normal: 9 value to 4 bluffs (missed this one pretty badly)
  • over: 2 value to zero bluffs

So this did quite well in this run, but buyer beware: if you try this yourself, expect high volatility. Doing this, you end up in a lot of big pots, and both with your bluffs and your value bets you’re going to lose a fair number of those. But that roller coaster is pretty exciting, and playing a high bluff style is a good way to start exploring what kinds of bluffs seem to work a little more often in various spots.

I’d also warn people that it is probably easier to go on tilt yourself playing in a style like this, and doing that for even a few hands will typically cost you quite a bit, so it can take a bit of practice and focus to play in a style like this successfully. I’d guess certain kinds of personalities are probably better able to do this, also.

So this one is probably not for everyone, but I have to say I found this to be more fun than say, the Passive Fish run a long time back.

2 Likes
190

One more observation here: if you are deliberately trying to have significantly more bluffs than value, you actually need to open pretty wide, or you’ll just wind up on the flop and beyond with too many value bets. I generally like playing fairly tight pre-flop, and so I needed to use a more LAG range in order to be able to bluff this often.

1 Like
191

OK, after playing in a very aggro style, I’m going to switch it up and play in the most passive style I’ve played yet on relatively higher stakes, 10k/20k. I’ll save the mechanics until after the run is over, but I’ll be naming this one: In Pairs We Trust.

192

In Pairs We Trust is being beat so far by the players at 10k/20k so far. I think I’ll try to tweak it a little, probably trying to get a bit more aggressive post flop. I intend to continue the ultra passive pre-flop lines, as this is really a remake of Lazy Limper, with a few new twists (it’s quite a bit tighter pre flop, which might be back firing).

A lot of very high ranking players on this site seem to be able to do quite well playing quite passively pre-flop, and here I’m trying see if I can’t come up with something that emulates that. It’s actually perversely satisfying to see myself getting destroyed playing this way… LOL… I probably need to get more fully behind this approach, and think harder about how I can make it work.

I’ll have plenty of time to think about it, as I won’t be able to play for a few days.

1 Like
193

OK… I think that might conclude my test runs for a bit. It was hard to stop while I still hadn’t lost a single session, but In Pairs We Trust managed to take care of that particular problem.

As mentioned earlier, this was a variant of Lazy Limper, which proved very successful in some low stakes runs. Here, I continued the tradition of no pre-flop open raising. I’d intended to 3-bet with almost any hand I was limping with, and so I was playing almost everything as a trap, but barring a few cases in very short handed play, almost no one ever raised over my limps (which was a source of some misery).

So all of the hands below were played for a limp (notice I’m playing every pair from every position, and intend to not fold any of them to a single, normal sized raise from any position… hence the name):

  • 6: all pairs, AKs - AQs
  • 5: AJs - ATs, KQs, AK
  • 4: A9s - A8s, QJs, JTs, AQ
  • 3: A7s - A6s, BWs, AJ
  • 2: A5s - A2s, T9s, 98s, AT
  • 1: SC, A9, KQ
  • 0: S1G, A8, KJ

The general strategy: play very carefully with normal hands, and wait for monsters. With monsters, bet them in a way that seems likely to maximize average return (so smaller bets on boards not likely to have connected with others, where everyone is showing weakness, and bigger bets when you think you have a fair chance of getting calls). I hadn’t initially planned on bluffing very much, but after really getting crushed over the first 500 hands, I add both a few steels of unwanted pots, and playing some of my draws like my monster hands.

What went wrong:

  • I ran a little bad, but I think a losing result for this feels fair, at least as I played it
  • Waiting for monsters works better at full ring than shorter handed, and I almost always had to initially create tables, and so probably 850 or more hands were with 6 or fewer players, and maybe 600 or more with 4 or fewer. I eventually just mostly ignored the opening ranges above when heads up or 3 way, though I mostly stuck with the passive theme (though I’d let the fireworks go with the rare players that would play aggressively)
  • I think the ranges above are just too tight, especially when you are only limping… I’d have been happier with it if the tables were more aggressive, but maybe the biggest problem of all here: I was playing too much like most everyone else, and so was doing less that gave me any kind of edge over anyone else

I do think something like this could work at the high stakes tables, and just being more careful to only play at relatively full tables would have gone a long way to help here.

In Pairs We Trust result:
Tables played: 10k/20k NL Holdem 9 max and 6 max
Hands played: 1,003
Chips won: -$2,951,636 (does anyone know how to make red font here?)
BB/100 hands: -14.71

194

Results summary:

  • Pre-Flop Hammer @ 1/2: 222 BB/100
  • Super Nit @ 1/2: 856 BB/100
  • Passive Fish @ 2/4: 73 BB/100
  • Lazy Limper @ 2/4: 372 BB/100
  • Value Village @ 5/10: 220 BB/100
  • Pre Flop Hammer @ 5/10: 241 BB/100
  • Value Village @ 10/20: 176 BB/100
  • Lazy Limper @ 10/20: 244 BB/100
  • Pressure Cooker @ 25/50: 139 BB/100
  • GAG @ 25/50: 300 BB/100
  • Robo TAG @ 50/100: 131 BB/100
  • PFP @ 50/100: 130 BB/100 (guest run)
  • GAG Cook @ 50/100: 190 BB/100
  • Pre Flop Hammer @ 100/200: 2 BB/100
  • Get da Fishy @ 100/200: 104 BB/100
  • LAG Problems @ 100/200: 125 BB/100
  • PFP @ 200/400: 30 BB/100
  • Orphaned Pot Adopter @ 200/400: 49 BB/100
  • Value Village @ 500/1,000: 109 BB/100
  • Balance @ 500/1,000: 153 BB/100
  • Balance V @ 1k/2k: 11 BB/100
  • The Maniac @ 1k/2k: 39 BB/100
  • Light Rain @ 2k/4k: 14 BB/100
  • Random Fool @ 2k/4k: 26 BB/100
  • Multiple Personality Disorder @ 5k/10k: 66 BB/100
  • Balance B @ 5k/10k: 109 BB/100
  • In Pairs We Trust @ 10k/20k: -15 BB/100
  • Small Ball @ 10k/20k: 3 BB/100
  • Super Nit @ 20k/40k: 58 BB/100
  • Laggy TAG @ 50k/100k: 27 BB/100
  • Night Rain @ 50k/100k: 77 BB/100
195

I’d been running dry on ideas for new runs, and hadn’t felt like re-running any of the prior tests, but I’ve recently thought of something similar to the Balance runs, but where the idea is to track aggression versus passivity on each street, rather than value versus bluffs. Here again it is easy to go off the rails on one side or another, and so I think trying out various relatively balanced targets can be nice.

I haven’t come up with a name for this yet, but think I’ll give it a spin in the next few days at 20k/40k, keeping track on each street of passive lines (checks and calls) versus aggressive lines (bets and raises).

1 Like
196

OK… some difficulties deciding how to set this up.

First: how to track aggression? I was first wanting to keep track of a lot of data: street, players still in, bet size, and check/call or fold if not a bet. Some people (or HUDs) track aggression frequencies by just the ratio of bets to calls, but that means a player that bets 3% of the time and calls 1% looks just the same as a player that bets 60% of the time and calls 20% (so that the ratio doesn’t tell us much about the strength of the range making the bet or call). Likewise, there is an issue with checks: is it really passive to check raise, for example? If anything, a check raise seems more aggressive than a donk, and so recorder the check as a passive action in that case feels a touch misleading. But by keeping track of all of the actions, people can build their own preferred formula for measuring aggression, or also to make several different ones.

I also felt that these ratios naturally change depending on the number of players still involved. Against a single player, you have a much better chance of having the best hand, or of being able to get a fold with a bluff, than you do with 8 opponents. So the ratio will naturally change in multi-player pots versus heads up.

Further, there is a big difference between someone constantly making small blocking or probe bets, and someone frequently over- betting the pot. But as I started putting all of these together, it made it increasingly difficult to decide on targets.

In the end, I started testing out a version, Bluffing Is Wrong, a test run with no bluffs allowed, just to see what a baseline aggression rate, for me at least, might look like. To continue applying at least a little pressure, I wanted to try and make a lot of thin value bets, and also add a lot of smaller bet sizes to allow hands like A high or even K high to be bet for value in at least some spots (again, just searching to expand spots where I’m betting, and hoping to get called more often by worse than by better hands). I’d also check some fraction of my value hands (some of the time with any value hand, but at a somewhat higher frequency with hands less vulnerable to being out drawn) to protect my checking range, even though this works against the general intent of being as aggressive as possible with zero bluffs.

Where am I stuck:

  • Just deciding on how to keep track of all of this while I play, while still multi-tabling (and not missing recording any actions) has been harder than I expected… I’ve tried out several formats now that haven’t worked out to my satisfaction
  • With any style where you are playing in an overly exploitable fashion, I prefer full ring, where you can kind of play a tighter range and hide in the shadows. Here, with Bluffing Is Wrong, my speculative hands (think stuff like 76 suited) have less value if I’m never going to bluff with them when I miss, and so that forces a bit of a tighter opening range. Both don’t work as well short handed (always value betting and being overly tight with your pre-flop ranges). So I decided I wanted 9 max, slow tables… but at 20k/40k lately, those just don’t fill up like they used to when I create a new table. So for this test run, I’ll probably move down to lower stakes where we typically have standing slow full ring tables running.

Anyway, once this is run, I’m hoping to have a kind of baseline that can be used as a point of comparison with other runs where I start expanding the spots where I bet:

  • range based c-betting
  • semi-bluffs
  • attacking weak or capped ranges
  • MDF frequency error exploits
197

Did want to mention that I do still intend to get back to this at some point, and eventually play at least 2 or 3 more of these runs. I’ve also enjoyed the suggestions people have provided, both on this thread and privately.

2 Likes
198

Darn… I guess I’m giving up on Bluffing Is Wrong. I’ve run into two opposing problems, related to the difficulty of tracking all of this info while multi-tabling:

  • I don’t want to play it short handed, as then I have too many hands to record, and so I want 9 handed tables
  • The only spots that have slow 9 handed tables don’t have a lot of them open, at least down as low as medium stakes

While I didn’t keep track of win rate (as I tried this at different stakes, and in several over several different periods), I think I lost money in the hands played, as I ran into a few bad beats, and also didn’t hit boards very hard (and without bluffing you just need to hit something with most hands you elect to play).

Just looking at statistics from my aborted session, I’m amazed at how passive the play seems, especially at lower stakes when multi-player pots are the norm. If you are only betting value, having multiple opponents really just reduces the frequency of your value betting. As an example, in the last session, on the flop, I bet 6 times (mix of half pot and pot only on the sizing in that run, for simplicity), but checked 27 times, folded 10 times, and called 5 times. In all of those, I was heads up only 7 times.

So if I do come back to this, I think I’ll need to do some things to manage the difficulty of taking all of the notes: perhaps giving up on multi-tabling (which then means the run takes a lot longer to complete), or perhaps playing super tight ranges to reduce the volume of hands played (which then skews the statistics toward aggression, as you have hands with higher equity).

So for now I think I’ll pull the plug on Bluffing Is Wrong, though I do still like the idea of doing some runs that focus more on the ratio between aggressive and non-aggressive lines.

199

OK, I have another run going at 50k/100k that I will be calling Hard Rain. I’ll avoid revealing any of the meaningful mechanics yet, as I think there are plenty of players at this level that read this thread and also are good at making adjustments, but suffice it to say that a few players have already commented that I seem to be more aggro than usual of late.

This is another where I’m trying to track aggression frequencies, and I’ve hit on a few simplifying features that is making tracking a little easier:

  • I’m only recording my final action on any street, and so if I check with 6 opponents, then raise over 4, and finally fold with 2 left, all on the same street, that’s just 2F in the section for that street
  • I’m recording bet sizes using my normal m (min), s (smaller than half pot), h (half), n (between half and full), p (pot), and o (overbet)
  • Passive actions are X (check), C (call), and F (fold)

I obviously lose some information by eliminating any preliminary actions on streets, but I think it makes for simpler evaluation later, anyway.

So far I’m down chips, a little under half way through the run, but not too bad. I’d say I’m running bad, having lost a with top set (AA hole cards) on the river, and have seen things take a sickly twist on the turn and river with a number of other really big hands, but honestly it’s often hard to tell if it is mostly because you are running bad or playing bad… I’ve certainly also picked a few poor spots for big bluffs.

200

Yay! First run finished in quite some time, and first of a new type of run.

Hard Rain results:
Tables played: 50k/100k NL Holdem 9 max
Hands played: 1,002
Chips won: $42,497,890
BB/100 hands: 42.41

Mechanically I was just playing my normal game, but putting effort into staying alert to reasonable spots to bluff on turn and river, and also some thin value spots, in an effort to play in a style that was at least slightly on the aggressive end of the spectrum, even if it was fairly timid next to what you’d see from gg or El-Jog.

I also generated a new c-betting algorithm, somewhat similar to what I used before in, errr…was it Laggy TAG possibly?

  • paired board worth 10 points, and trips on board worth 15
  • subtract 10 per opponent beyond first
  • A: 40 points
  • K: 30 points
  • Q: 20 points
  • J: 10 points
  • 5: 5 points
  • 4: 10 points
  • 3: 15 points
  • 2: 20 points

The cumulative total of points then become seconds. If 60 seconds or more you bet. If 30, you bet if the second hand is between 12 and 6. If 15, between 12 and 3, etc. At 30, you will make a half pot bet. At 60 or more, where you are betting with 100% of your range, you will make a 20% pot bet. And if you bet with 5 or fewer points, you will be betting at least pot. Additional opponents drive sizes down. Coordinated boards and situations and possession of vulnerable value hands will drive sizes up.

I used the second hand of a battery powered clock that was not synched to the internet, ensuring that even if people knew what I was doing, the daily drift of the clock away from perfect time acted as a further randomizer.

I was reasonably consistent about using this with 1 or 2 opponents on the flop. With 3 or more, I would often ignore it, and just make normal decisions, though I might still look at the clock to decide between more aggressive and more passive choices.

I’ll return shortly with the interesting part (as soon as I organize them): the aggression statistics.

201

Oh, brief mention on opening ranges used in Hard Rain: tighter early and wider approaching button. Ranges typically looked reasonably like the Laggy TAG ranges, though I was wider than than for the first 400 hands or so, and finished tighter than that, just to make the burden of note taking a bit easier. So in general I had a tight set of hands I would always play from a given position (here a bit wider than Laggy TAG), and then other hands I would play based on the “aggro meter” (the second hand of my battery powered clock).