I been thinking on the balance of the best player in terms of predicting your opponents moves.

Bare with me while i explain my point of view: I personally fear playing frank a lot more then playing Gnuffone, szew, Gui or dead piggy (which are/were ranked higher on ladder). This might just be due to my personal game historic ofcourse. I am not trying to challengE the skills or merit of any of the players i mentioned, just using them as references for the point i'm about to make.

I have this deeper fear of running into frank because most of the times i lost against frank, on a well focused game, he largely beat me by predicting my unpredictable behavior.

In my opinion unpredictable behavior is what gives you the edge on a well balanced game. Some moves are highly predictive because they are the most rational safe way to victory, but when they are predictable they stop becoming the rational safe way, you could be running into a trap, so when you go against people who are used to predicting the best possible opponent play you start playing this bluffing mind game, where it starts making more sense strategically to play unpredictably: Flanking with your stack and attacking the main with lower number instead of the opposite. Minimal attacking with 3 or 4 instead of 2 because your opponent could deploy minimal. Delaying like hell to make sure you get the last action to break or finish an area bonus. Attacking against a huge stack you know might no longer be there when you attack. Attacking with 80% so your opponent won't get the area once your stack is gone, ect.

These are plays that might or might not pay off, it all depends on how well your opponent predicts your unpredictable behavior. And i guess this is where you really play the player and not the game.

My question is: Where do you draw the line? At some point the opponent knows that you know he knows you know that he knows you know his next move. So you have to go back and play "solid". From what i've seen the higher top ranked players play mostly solid, and try not to go into the bluff game that deeply. Opinions?

The best way to ensure you stay unpredictable is to roll a die or flip a coin. Humans are actually pretty bad at behaving randomly, no matter how hard you try.

At some point the opponent knows that you know he knows you know that he knows you know his next move.

What?

I do think that there are a certain number of moves you can do in a certain situation, which doesn't throw the game away. Being unpredictable is playing the one which they think you won't, from what they can see and how you usually play.

If you are in a losing position, then play unpredictably, if you are in a winning position, play predictably and secure the win without doing anything stupid. Also you can never know if they will know what move you are going to do, depending on how much they have thought through and what is normal and unpredictable play for them

Playing warlight is a little like poker

if you are playing a new person you need to play solid making the best choices you can. You cant assume a person will opt for the 2nd best move to bluff you otherwise you will lose out more often than not

if you are playing someone you have played many times before thats when you get to know your opponents style. That's when the bluffs and double bluffs come into play
you should fear these games. you should relish the challenge ;-)
- if you know your opponent has seen you play the obvious move time and time again and he is predicting your moves then its time to mix it up a little.

I think that usually there is an intel (or other in-game) advantage of some sort, which helps (one player or the other) with predictions or with the mind game. It is a bit rare for both players to have the same amount of intel. Taking risks are easier when they are hedged in some way or if it is a leveraged risk (leverage based on your intel, positional, other advantage).

I base my predictions on intel, the player's tendencies, positioning (who should want what), game flow (each deployment, move, move order is a hidden language of some sort), my tendencies (what have I done lately? should/can I slightly or radically change what I am doing mid-game to get a better result?).

For 1v1s, I think there are two key factors involved.

First, as ps mentions, some styles of play are just naturally difficult for your own style. I think this is where 知彼知己 百戰不殆 (know the other, know yourself; 100 wars without losing) comes in. Knowing the enemy's style/tendencies is one thing; but to adjust properly requires self-critical knowledge of your own style/tendencies.

Second, it is easier to understand how someone below your level of play thinks. It is harder to know the thinking/style(s)/tendencies of someone more or less at your level or better. And if their style naturally causes problems for you (first factor above), the problem could be worse. What I like to do to help deal with this: play the player in question in a few 1v1s; don't do anything special, just feel the player out at vanilla (basic style, nothing tricky/special). While playing, don't worry about winning or losing. Just try to get inside the other player's mind to try to understand why he is doing what he does. Then, review the history after you lose (if he is trying his best and you are just trying to understand him, you should lose) to gain some more thoughts/ideas. Greater familiarity helps.

ps I know exactly what you mean.

About a dozen players are *good* at predicting me. Much better than I am at playing or at being unpredictable. This has the effect of them beating me much more than I beat them. Does this make them better than me?

You mention "solid moves". I learned most of my technique on the ladder or autogames with randoms (as most of us do). I was playing people of medium to high skill, so I learned to play moves that work well against okayish players. We call these moves solid, but actually theyre extremely defensive. This has taught us to consider a position won when it is infact not.

One way to get around humans inability to be unpredictable is to immitate another top player; flaws and all. I have looked at a lot of sze, zaeban timi and luxis games. They each have a different attitude to the game and if I am playing someone who predicts me I can try to wrong foot them by picking all the 4 territory bonuses and then getting in their face like zaeban does. Or guessing their picks and countering them like timi does.

I find flipping a coin isnt actually that helpful. It makes me unpredictable, but not in a way that wins games. It is useful when there is a fork in the road where there are exactly 2 options and winning is about picking the right one. How often does that happen?

What I was trying to say about "solid moves" is... They're often not the best move. They get best results against okayish players, but thats because okayish players will eventually hang themselves. The best move is usually quite risky, even when youre winning.

you're assuming the other player always plays the same way. i don't think i can fully agree with that premise. although there is a certain level of knowledge on game mechanics and personal confidence that you can assert, some of your opponents key decisions are usually made on gut feelings, and often some of those decisions are made against their own better judgment. so even if you know the player 100% you should never truly, completely, be able to predict his behavior at all times.

which makes the whole matter even more interesting.

for example i really love when there is a battle on 2 different spots of the map at the same time, one of them with higher strategic value and you have to guess which one your opponent will try to wipe you out from. All sane reasoning says the higher troop investment will be on the higher valuable area, but that's such an easy prediction to make that sometimes you both end up battling harder for the minor area.

yeah, i agree with you dead piggy, on the matter of "solid moves"

Dirty cleanliness.