Something I do two or three times a year is analyze how my starting hands are faring in different preflop situations. For example, I look at how hands I’m raising first in from the button, cutoff, hijack, under the gun, and the small blind are faring. I look at how hands I’m defending from the big blind fare against opens from different positions. I also look at how hands I 3-bet from the small blind fare against button and cutoff raises. Etc, etc. These are all situations where I have a fairly substantial amount of data to work with and feel confident in the results. Based on the analysis I add or subtract hands from the various ranges. This process has contributed a great deal to my improvement over the past few years.
How important is preflop play? There are 169 starting hands in hold-em. 13 are pairs, 78 are suited hands, and 78 are offsuit hands. There are 12 ways to make an offsuit hand, 4 ways to make a suited hand, and 6 ways to make a pair. You’ll be dealt a specific pair like Aces once every 221 hands, a specific suited hand once every 332 hands, and a specific offsuit hand once every 111 hands. More on the math associated with all this is here
Now let’s assume the following plays are mistakes and that you make them:
Raising K2o, K3o, K4o, K5o from the button against typical blinds
*Not* raising Q8o, Q7o, Q2s, Q3s, Q4s, Q5s and Q6s against typical blinds
Let’s assume you lose the following when you play or don’t play these hands. These are approximations based on my data:
K5o (.03) BB/HD
K4o (.06) BB/HD
K3o (.09) BB/HD
K2o (.12) BB/HD
Q8o (.15) BB/HD
Q7o (.05) BB/HD
Q6s (.09) BB/HD
Q5s (.07) BB/HD
Q4s (.05) BB/HD
Q3s (.03) BB/HD
Q2s (.01) BB/HD
Last, let’s assume you’re going to play 1000 hands of 3-handed poker. That means over the course of these hands you’ll be on the button roughly 333 times. So how much do these mistakes cost on average over these 1000 hands?
Since you’ll be dealt a specific offsuit hand once every 111 hands on average you’d be dealt each offsuit hand 3 times on the button and lose:
.03*3 + .06*3 + .09*3 + .12*3 + .15*3 + .05*3 =.09+.18+.27+.36+.45+.15 = 1.5 Big Bets
And you’ll be dealt a specific suited hand once every 332 hands on average or about 1 once on the button and lose:
.09+.07+.05+.03+.01 = .25 big bets
All that adds up to a net loss of 1.75 Big Bets over the 1000 hands. That might not sound like a whole lot but consider that’s .175 big bets per every 100 hands. Now consider that some of the best limit hold-em players average 1 big bet per hundred hands, that’s 17.5% of their profit! Also consider that these are a just a few of hundreds of possible mistakes people can make. Imagine they are 3-betting incorrectly from the small blind and defending incorrectly from the big blind over this 1000 hand sample and they could be giving up quite a bit.
Nobody plays perfect preflop, it’s nearly impossible. But the idea is to make as few mistakes as possible. The best way you can do this is to analyze your own data and see how you’re faring with hands in certain situations. This should give you a general idea of what hands are playable under normal or average conditions. You then can make in-game adjustments based on table dynamics. For example, if you have very tight blinds, you can add in all those hands that were small to marginal losers from the button. If you have loose, aggressive blinds, you can drop all the hands that are small winners. Another example is defending from the big blind against the various positions. I know what my default defending range is based on a typical raiser’s RFI% from each position. If the raiser is atypical, I’ll cater my range to them. I also account for the player’s postflop ability when deciding to defend borderline hands. In extreme cases, I’ve played against opponents who play so poorly postflop that I defend 100% of my hands from the big blind regardless of what position they’re raising from. So use your data as a baseline and go from there. It’s not an exact science and you won’t get every situation right. The most you can do is to consider all the information available and make the best possible decision based on it.