We were talking about some of the crucial tips that start beginners on the road to victory, and trying to focus on the ones that people, myself included, often forget to mention. The first, and most critical of these, can find its roots in the old chess adage, "The best form of defence is attack." In short, rather than being held in supposed defence, as many men as possible should be tapped sideways.
When the game seems to be tilting out of your control, it is the knee-jerk reaction to hold you creatures back to stem any further bleeding. However when you do this, you will often be sacrificing your only chance of winning the game. Let's say you are being attacked by an Aven Windreader and there's nothing you can do about it. The only other cards on the board are an opposing Fugitive Wizard and your Hill Giant. Here, you have the option of holding back the Giant to prevent a further 1 damage a turn or you can attack.
There are two concepts that state you should be attacking, and the first is the ratio of damage. By holding back the Giant, you are taking 3 a turn and dealing none, whereas by attacking you race 4 points for 3. The second ratio is obviously far better for you (assuming the starting life totals are similar; if that 1 damage costs you a turn, then attacking back is not worth it).
The other idea you should be thinking about is what happens if you draw the removal spell you need. Look at what the game state will be like in both cases of do or do not attack. If you held back, this might, in the long run, gain you an additional draw to your removal spell. There are several things of note that happen when you do draw it. It might arrive in time turn the game around, or you may be too far behind and on too little life for it to matter still, if your opponent has drawn other threats.
By attacking, if you draw more creatures than your opponent, you might be able to swing the race in your favour such that either it is now he who needs the removal spell, or it might be that he will now have to hold back the Windreader to block with, effectively turning your Hill Giant into the removal spell you had been so desperately looking for. Secondly, and most importantly, had you attacked each turn, now the removal spell, instead of just digging you out of a hole, would put your opponent in one as he now needs to find a solution to your Hill Giant. By attacking, you make so many more of your draws live. If you have a Lava Axe, it suddenly becomes an out too; even a Reviving Dose might be enough to swing the race in your favour for a turn. More creatures will also tilt it in your favour.
Attacking also creates the illusion that you are winning. Attacking portrays confidence—be it confidence in the current game state, in a trick in your hand or what you will do after combat, you are sending signals that you are happy with your attack, happy with the blocks your opponent can make and happy with what the game state will look like by the end of you turn. This confidence is a very powerful thing. It will force your opponent to make different plays, to block when he would not have otherwise done or to put you on certain tricks that you may not have. More often than not, he will just take the damage, which will be free damage for you that you would not have otherwise had. It should be noted that if your creature was intending to block (this does not go for walls or 2/4-esque creatures), then it probably should be attacking instead.
Another thing attacking will do is give your opponent more opportunities to make mistakes. The mistakes could be as simple as taking the damage that they would not have otherwise taken (although this is more often simply the correct play) or to chump block out of turn, or to put you on a specific trick that you do not have and then play around it all the time, or to just walk into a trick that you do have. By making the opponent face more decisions, you invite him to make more mistakes.
Jeroen Remie once commented that he hated creatures like Nezumi Cutthroat, which forced the person playing it to play it correctly. Had it been a Neurok Spy or Phantom Warrior, there was always the chance that an opponent would decide to hold it back and block or trade it on the turn it was made, but Cutthroat never gave them the option to make a mistake, and turn after turn the Cutthroat would relentlessly, and correctly, attack.
It is not always best to be attacking. For example, if you are on the back foot and by not attacking, you take no damage, or take less of a damage ratio than had you attacked. And it is often not correct to continually bluff attack. In order to bluff attack, you have to convince your opponent that your attack is credible. His potential blocker must be worth using your trick to kill and your attacker must also be worth saving.
Tim's maxim was to get beginner players to turn every creature sideways, every turn. This obviously has some problems if your men are just outclassed, but it is a good place to start. To try it out, I drafted a Time Spiral Block deck, forcing white-green, where I took no one-power creatures (because they have little use if they must attack) and every pump spell I could get my hands on, so that my men could survive and that it was credible for me to continually all-in. I wound up with a very bad deck featuring Gift of Granite, Healing Leaves, Wrap in Vigor, Dawn Charm, Aether Web, and Fortify for tricks, and an unexciting bundle of creatures.
In the first round, it was completely correct for my men to attack every turn and I would have done it even if I were not playing to the maxim. There were only two occasions in the second round where it was normally incorrect to attack but in one I had the trick to keep my guy alive and in the other, my opponent chose not to block, meaning attacking was still the way forward. Unfortunately, I was destroyed in the final where my opponent's much faster deck made a mockery of all of my Norritt-loving creatures, because I was too far behind and should have been focussing on staying alive rather than trying to win an impossible race.
It was an interesting experiment, especially given the success it had. One of the examples I want to look at was this: My opponent has an untapped Thallid Germinator in play and has a freshly played Aven Riftwatcher. I have a Forest, a Plains and a Spinneret Sliver. It is my fourth turn and I have already missed one land drop. The attack maxim means I have to get my Sliver in the red zone. It is an interesting case, because a bad player will always block here without giving the situation much thought because both the Sliver and the Aven are not worth saving with tricks and all the tricks I could use to trump the Aven are worse than the Aven itself.
Another mistake new players make is they use their removal on pointless targets. They have some kind of compulsions to throw their removal at the first thing that moves or mildly upsets their plan. A creature is only worth killing if it really screws your plan up. Often it will just trade with one of your guys and you will continue to beat with the others. It is pretty easy to identify the must-kill targets because they are either a card you cannot deal with, an amazing creature or something that significantly turns the tables so that you are now winning.
As a simple rule of thumb, assuming you have a favourable board position to start, if a creature still leaves you with an advantageous board, even if it is only slightly in your favour, then you do not need to kill it. You should kill things that only tilt the board in his favour, and even then, only if it massively tilts it. It's ok to be slightly behind (say, beaten up by an Ashcoat Bear for the first few turns of the game, because on turn four or five the Bear will become meaningless as it will then be dwarfed by your guys).
Something that might be worth trying is practicing restraint. Now I'm not talking about religious morals or anything of that ilk, just resist the first impulse to kill something (unless it has a sign on its head so big it might as well be a bull's eye). Wait and kill the second creature that comes along, unless the first leaves the game under control. You can also just wait a turn to not kill a guy, to see if your deck coughs up a better solution or his spews up something more deadly.
I see players trying to play around opposing combat tricks all the time. They keep refusing to block, taking more and more damage, until several turns down the line, they are forced to block. When this happens, one of three things will occur, and none of them are particular good. They will either have had the trick all along, beat you with it now and have dealt you all the damage in between; or they will have drawn the trick you feared they had whilst you gave them all the free time and wreck you with it now; or they never had it and you took all that damage for nothing.
It is always a fine line as to whether to block or not, but there are a few simple guidelines that will help you in making a decision. The first thing to think about is if you do not block, is there anything else you can do? If your hand contains nothing that will change your blocking situation next turn, then it might be in your interest to block. When is it good for you to block? You should block if you (a) do not mind simply trading creatures, (b) do not mind trading your creature for the trick you suppose him to have (it is normally in your interest to do so), (c) you have no real option but to block because you are low on life, or (d) your creature is expendable and you wish to find out what the trick is.
If you have a pump spell of your own or an instant removal spell in your hand and you are tapped out, it is normally correct to not block (unless trading your creature for either of his cards is good for you). This is because next turn when he repeats the same attack, you will have the flexibility to respond to what he does if it swings the board parity against you. It goes without saying that if you had mana up and a trick the first time, you should also block (if you want to).
If you do not do any of the above, try applying them to your game, maybe in extremes to start with if the theories feel unnatural to you, so that you see the strengths in what I am suggesting. From there, you can tailor the extreme to the good and continue to progress as a player.