Late position hands when first in

Rather than itemize subsets, I'll go ahead and post results en masse of large-scale trials, and then itemize key points afterward.

The Maverick profile on Turbo Texas Hold'Em plays the following hands in late position when 1st in:

AA-44, AKs-A2s, AKo-A10o, KQs-K2s, KQo-K9o, QJs-Q5s, QJo-Q10o, J10s-J7s, J10o, 10-9s to 10-7s, 10-9o, 98s-96s, 87s-86s, 76s.

I tested each hand over 100,000 trials, with the button locked one seat to the test profile's left, and with no other players having entered the pot.

Due to the high volume of playable hands in late position that I needed to test, hands that would obviously turn a profit when raised were omitted. There's no need to run 100,000 trials for open raising AA or AKs, when we know those hands turn huge profits in most situations, let alone when 1st in. Several other hands that showed a healthy profit in middle position were also omitted.

All hands 1st in opened with a raise. We've seen in previous trials that, with few exceptions, hands show maximum profitability when 1st in with a raise over limping when 1st in. As you find later position preflop, raising becomes even more important, as you vastly improve the chances of folding marginal hands behind you and that only players in the blinds will call, often with lighter holdings than they'd typically play against a raise due to having blinds in the pot and receiving a discount to call. Thus you not only will usually have the best position in the hand postflop, but your holdings on average will be stronger than theirs.

Pairs	Net $ Per Hand (3/6)
88	0.90
77	0.25
66	-0.12
55	-0.57
44	-0.85

Aces	Net $ Per Hand (3/6)	Aces	Net $ Per-Hand (3/6)
A9s	1.20	A5s	0.26
A8s	0.70	A4s	0.15
A7s	0.47	A3s	-0.01
A6s	0.19	A2s	-0.06
A10o	1.11	A9o	0.39

Kings	Net $ Per-Hand (3/6)	Kings	Net $ Per-Hand (3/6)
K10s	1.08	K3s	-0.67
K9s	0.52	K2s	-0.70
K8s	-0.01	KQo	0.68
K7s	-0.20	KJo	0.70
K6s	-0.32	K10o	0.31
K5s	-0.47	K9o	-0.33
K4s	-0.61

Queens	Net $ Per-Hand (3/6)	Queens	Net $ Per-Hand (3/6)
QJs	1.15	Q7s	-0.60
Q10s	0.84	Q6s	-0.63
Q9s	0.27	QJo	0.51
Q8s	-0.30	Q10o	0.12

Others	Net $ Per-Hand (3/6)	Others	Net $ Per-Hand (3/6)
J10s	0.70	J10o	0.00
J9s	0.10	10-9o	-0.74
J8s	-0.30	98s	-0.48
J7s	-0.70	97s	-0.84
10-9s	-0.02	87s	-0.80
10-8s	-0.32	86s	-0.98
10-7s	-0.83	76s	-0.97

Pairs: You can profitably open with pairs down to 77. 66 and any lesser pair posts a loss.

Suited Aces: Even the marginal A9s posts a big profit when open raised. The other aces show progressively lesser profitability until we see a small loss with A3s and A2s. It's possible the smallest Aces can still profit if there are enough callers in the hand due to their drawing status and the need for great pot odds to draw.

Ace-10 offsuit is a big winner. A9o also showed a profit, though the drop was precipitous enough that A8o was figured as a loss (a 100K side trial confirmed that A8o posted a -0.12 net).

Kings: K10s and K9s posted a reasonable profit, but K8s posted a small loss, and losses mounted for the remaining suited Kings. Offsuit King-broadway hands posted a win, but K9o posted a sizable loss, over $0.60 per hand less than K10o.

Queens: QJs posted a sizable gain, as did Q10s. Q9s even showed some profit when opened, but Q8s showed a sizable profit drop and a loss, with subsequent lower hands posting big losses. QJo posted a significant profit, with Q10o posting a slight profit, indicating a small margin for error in playing such a hand profitably.

Lesser hands: J10s showed a sizable profit and J9s showed a slight profit, but J8s and J7s showed losses. No ten high suited or lower suited hand showed a profit. J10o broke even and 10-9 offsuit posted a big loss. The drawing power of these hands do not make enough profit against the rake in 3/6, so they aren't playable as they would be in many other limit hold'em games.

Using this data as an indicator, a revised preflop open raising strategy would raise profitably with these hands:

AA-77, AKs-A4s, AKo-A9o, KQs-K9s, KQo-K10o, QJs-Q9s, QJo-Q10o, J10s-J9s

This makes up a bit over 18% of all possible hands.

Ill finish this entry with a digression: Since this was tested in a game with 3-4 players to each flop, a $3 max rake and $1-2 jackpot drop, and a little under half of the hands played being raised before the flop, obviously some adaption would be required for looser games.

But on average, 3/6 games in many areas often feature the aforementioned sub-optimal environment, and that's the main goal of this project in any case: to devise a winning strategy in a sub-optimal 3/6 game. We already have strategies designed to succeed in more ideal low limit games that work when 6-7 players see a flop in a passive game. Small Stakes Hold'Em was written exactly for these games, and the strategies in most books are designed for these games. When it comes to more sub-optimal games, basically every writer recommends that players not play in these games.

Most of us don't have the luxury of being able to headhunt for great games, thus these efforts.