The thread that used to be known by the name "The biases we know and practice"

[Ghost of the dead horse]

The thread that used to be known by the name "The biases we know and practice"

I have one; I always believe I can program some difficult things in 1/2..1/10th of the time it actually takes. I wonder why I fail estimating the correct time. This is also when I try to take into account my biases and then think I now have good guess.

I just played a team game of warcraft with a teammate who shouted in the start, "rush = win!". I hate the lack of subtlety and ability to manage other plans.. especially when that particular tactic is too obvious and not good enough a battle plan. Expecting good outcome from little to no reasons shows gross incapability of judgement. For you computer game speak illiterate, the expression ment a tactic where we gather all our forces as fast as possible for one single attack, which "should" produce a win.

So I asked how many games he lost with that tactic and lost: it was 100. So I asked why he is so sure of it? No reply. Why can he be so sure? Answer: he's just 99% sure. And no, he has not played 10,000 games. So where's the certainty he gets? I asked if the opponent goes by the same tactic. And he thought that they do. I asked why they don't win then? Ok at the point I was asked to STFU and I understand, I normally never talk so much in a game.

I don't know if he saw the error in his thinking. But after so many losses and practicing the most common tactic in the game, with opponents surely practicing the same tactic in most of the games, he should have grown out of the bias that it is an almost sure tactic.

My women-hunting days had me think that every woman hold the chance to be the best ever, and I would find true love. No, I bored/bailed out most often, and I was sometimes dumped. I understood that I loved the excitement of new people, and I thought to have evaluated the person correctly every new time, having got rid of my bad ways (in whatever way they were).

What makes these biases? Is it in the belief that some issue is only a matter of mind, so that the mind can be bent to accomplish anything?

Is the key reason for this behavior in there, that optimism gets better results on average, so we decide to be optimistic, even tho when the optimism produces wrong results.. and that the combination of all of these things is too difficult to find out and correct.

If it's the matter about optimism, one could do best to outcome one's own internal biases about a situation he/she can't correct, while maintaining optimism in other issues. Why doesn't this happen either?

 

[ptgatsby]

I just played a team game of warcraft with a teammate who shouted in the start, "rush = win!". I hate the lack of subtlety and ability to manage other plans.. especially when that particular tactic is too obvious and not good enough a battle plan. Expecting good outcome from little to no reasons shows gross incapability of judgement. For you computer game speak illiterate, the expression ment a tactic where we gather all our forces as fast as possible for one single attack, which "should" produce a win.

Assuming you are talking WC3, rushing is generally the correct strategy in any non-1v1 game.

If you assume a norm of 50/50, I believe not rushing is around a 20% victory rate, whereas mutual rushing returns to a 50/50 rate (normally in 3v3+ games, the amount of time the 3v1 rush lives determines the likelyhood of it becoming a 2v1 game... but even in 2v2, defensive is highly inefficient).

I do not, however, read Korean and so the statistics at the competitive level are kinda rough for me. And it differs slightly by game and such (this was for starcraft, however a lot of the rush was "early pressure as soon as possible" more than all out rush, but that's because it was 1v1. Rushing was universal in 2v2 team tournaments, except in cases where terran tanks were available already.)

Instead of saying "rush = win", it should be said "A lack of speedy attack reduces our chance to survive their speedy attack while giving no benefit if they also do not attempt a speedy attack".

This is why early pressure is a core part of most strategy games (some exceptions may apply, such as Supreme Commander, large maps, etc... which normally don't involve containment or prevent early enough pressure).

What makes these biases? Is it in the belief that some issue is only a matter of mind, so that the mind can be bent to accomplish anything?

I think there are two biases that you are talking about.

The time bias is a bias in our ability to simulate the future. It is associated with how we are unable to understand extreme risks (small and big). After a certain point, the human mind begins to abstract the information into terms it can understand.

(For example, the human mind can only identify and associate so many "faces" in its memory. We have greatly exceeded this amount (use to be at the tribal level). This is true at every level - we can only associate so many as friends, so many as tribe mates and so forth.)

This also happens with planning. and predicting the future. We are one of the few species that can simulate the future, but we do so poorly. For example, pundits on TV that assign a probability to a future event are wrong so frequently and so dramatically that I would condone laws to prevent them from being on TV (ie: it can be as bad as 90% confidence level with 10% success rate...). This is true for most project plans as well. Only half finished projects that they associated a 99% confidence level in.

It is part a lack of ability to simulate future events and time, but it is also a lack of understanding of odds. The 99% came from drastically overstating the amount of time needed, and yet still had a very high non-completion rate.

I've seen it happen here at work. People reduce the amount of time it takes to do a task after enough time has passed (despite logging the time!)... and when faced with the same task, base it upon that memory, thus allocating too little time again.

(Good reading: Buehler, R., Griffin, D. and Ross, M. 1994. Exploring the "planning fallacy": Why people underestimate their task completion times.)

Is the key reason for this behavior in there, that optimism gets better results on average, so we decide to be optimistic, even tho when the optimism produces wrong results.. and that the combination of all of these things is too difficult to find out and correct.

No; least not in terms of cost-benefit. I think the general bias is just a sum of biological biases - smart in evolutionary terms, but not well suited for long term planning or risk taking. (We developped it to survive the hunting phase - simulate the value of a fire-hardened stick, for example... but not to simulate long term effects. Humans really are 'middle' creatures.)

 

[Ghost of the dead horse]

Yes it was warcraft. I play rt 4v4 most often, and that's the game style I ment. I see it differently, that the game favors creeping in the start because of the ease of base defence.

My point was in the belief that one can assure win with a tactic that everybody uses, whereas similar tactic and starting point from both parties would give you an expectation of 50%.

Oh I must go now, sauna time..

 

[Zergling]

Yes it was warcraft. I play rt 4v4 most often, and that's the game style I meant. I see it differently, that the game favors creeping in the start because of the ease of base defence.

My point was in the belief that one can assure win with a tactic that everybody uses, whereas similar tactic and starting point from both parties would give you an expectation of 50%.

Lots of people in computer games just seem to be idiots when playing for some reason, I guess this thread is to partially figure out why.

In terms of the "rush=win!" mentality, that may also come from a social pressure type of effect in computer games, where in order to appear better to other people in the computer game, or at perhaps be a bit scary to them, a lot of people will pretend to be better than they are. If that effect leaks into other associated areas, those people will take the bluster mentality into other interactions in the computer game, such as your setup.




In terms of time it takes to complete things, I really don't know very well where this would come from. Usually with me I seem to guess somewhere around half to twice the time to do most things, but the less I know about something, the more inaccurate it gets. I don't really know too well what is going on in my head for figuring out guesses on times other than some combination of "Other people took ... to this (for things I don't know well)", and "It has taken me ... in the past" for stuff I do know well.

 

[ptgatsby]

Yes it was warcraft. I play rt 4v4 most often, and that's the game style I ment. I see it differently, that the game favors creeping in the start because of the ease of base defence.

My point was in the belief that one can assure win with a tactic that everybody uses, whereas similar tactic and starting point from both parties would give you an expectation of 50%.

Oh, I agree... however the underlying part of this (in 4v4 on a 8 player map) is that no amount of base defense can even delay 3v1 (or normally 2v1). This makes the speed in which you attack the most important consideration. And most defense is inferior anyway, unless the home army is there to support it (full commital makes the rush strategy more powerful).

This is true in almost any game that doesn't have exponential resource growth (wc being limited to the gold that can be taken from a single mine).

I agree that if both follow the same strategy, it remains 50/50... but the point was that if you do not follow the same strategy, you open yourself up to a ~20% chance of winning if the enemy follows the same strategy. Hence, you need to have a pretty good reason not to follow the normal strategy (ie: the optimum strategy).

There are probably situations that work that way, but they require pre-made teams and good communication (and much practice).

Course, WC3 has other factors (creeps, etc), but it still tends to be true (I don't watch WC3 battles, where micro wins at the professional level... but at the amateur level on bnet, it plays out similar to SC).

 

[Ghost of the dead horse]

Perhaps I utilize some kind of statistical niche in the game. I score in the top 1-2% in the ladder and have scored top rank 10 in other ladders. On the average, I get my (random) team to help on base defence often enough that it works out.

In the start, the buildings (and militia, towers, moonwells etc) are stronger compared to attacking units, when compared to the middle game, where they become insignificant. It has just been my experience that so called "cc" tactic places me in the top 1-2% bracket, when I cc most of the time (70%). (cc = creep and counter, or play a reactive rather than proactive gamestyle).

I have found that in rt 4v4 games, I most often get enough of allys to help whatever base is under attack, but it is hard to convince everyone to attack in a coherent fashion. Perhaps it's because I am less of a force to force people to do things my way. I am almost the ultimate hippie and the liberalist in my rt matches. Defence is easier, and I get my team to defend. Attack needs more coherent battle plan and some ordering mentality; I don't get the people to attack so often, so I practice the rush tactic less. Often during play the people notice that I'm the best player and expect my help and start to expect my lead without me asserting myself as any leader. Another thing that happens is that the people tell me how much I suck. I notice that this is most associated to the originality of the choices I make, and secondarily to the outcome of said choices.

The ease of defence depends on the use of teleportal scrolls, which makes it easy to defend. I don't know. This has just been my observation. Oh and a point: after some while of playing, everyone starts to be able to utilize their teams better. So the team communication dependent aspect of the attack/defence choice becomes less pronounced, and you can expect better coordinated attacks late game, rather than in the beginning.

I have a huge known bias, or preference. I am the ultimate strategist and long term planner. I suck in micro, and I have what seems like a 300ms lag. So I'm inclined to use strategy rather than micro to win my games. After 3 seconds of back-and-forth microing 1 or 2 units I start to think, how could I better use my time? Does my town need management? Do I have to take a new mine? Should I see that our units are appropriate for what they have made, in accordance with the scout info I've got?

Also when recovering from the attack, you should consider attack/defence capability of each party and creeping capablitity (and profit gained) from each party before deciding on whether you should creep or counterattack.

and the level of early defence: 1 tower and 4 of your units vs enemy with 4 units, it is a tough match but turns into your favor. Even if it is a 2-5% difference, it's better to defend.

The reasons for choosing the attack tactic is that of lesser strategic ability and the fear that a superior strategical player beats you in the long run. In the beginning, you can harass well if you choose hero with a little more range and a bit higher dps. These are things you can learn easily. This is why rushes are done: people have as their ability, on average, reaction time and the ability to memorize - not the ability to strategize.

 

[ptgatsby]

and the level of early defence: 1 tower and 4 of your units vs enemy with 4 units, it is a tough match but turns into your favor. Even if it is a 2-5% difference, it's better to defend.

Perhaps, and yet the cost of 4 units is rarely (tempo especially) made up for losing even a couple of gatherers, which can be prevented in some cases, but not all (especially if the hero hits 3 or so). Worse, it limits your ability to move freely without some way of returning to base (one scroll rarely is enough). It is difficult to push once you are under constant threat at home. Generally, 4 troops at home = 4 units not producing much good. (Undead being an exception here).

This is coming from a heavily turtle myself, although I never played WC3 competitively, so I could be wrong about the game dynamics. My high level play knowledge is at the professional level where I watch games played by the top pros... a large gap between the two. I don't know the middleground well.

 

[Ghost of the dead horse]

Perhaps, and yet the cost of 4 units is rarely (tempo especially) made up for losing even a couple of gatherers, which can be prevented in some cases, but not all (especially if the hero hits 3 or so). Worse, it limits your ability to move freely without some way of returning to base (one scroll rarely is enough).
--
My high level play knowledge is at the professional level where I watch games played by the top pros... a large gap between the two. I don't know the middleground well.

I don't sit at base, and I encourage my team not to. There is a time you need to move in to your base to protect. This time is often considerable, so I use scout buildings, other scouts and general experience of game to know where someone may move. I have noticed that I am able to get teammates to often teleport at 20 seconds notice. This requires me to make 1 scout building (or other method of scouting) 10-15 seconds away from each base, so that I have a warning system. Meanwhile, the time can be spent creeping, or reacting to the attack. This is just statistically (observed) enough often enough to be viable strategy on average.

I try to start from the positions away from base, so that I can continue creeping near base after the game has progressed.

Also in the later game, the cost of portal scroll becomes insignificant compared to advantages gained in level. I make mines in my games and expand when appropriate and inform my teammates of such possibilities. This is something that I don't know most of other players to practice.

Yes it's amateur high level, and I wouldn't know how it becomes different with real pro matches. Perhaps I should step up to mid level, some arranged clan games.

Pre-arranged teams are also different from random matches. You trust your teammate there and you have better communication, and you have learned more of your friends so as to adapt to their playstyle. That's also what you can expect of your opponent.

 

[ptgatsby]

Interesting, I'd love for Bnet to record all the ladder matches so that a rough dataset could be constructed. Would be very interesting!

Anyway... back on topic

I guess I can flip the question around - do you think your strategy is universally better, or do you think you are exploiting a particular hole in the common rush strategy?

 

[Ghost of the dead horse]

I am not sure about if my strategy is really universally better. Instead, the game favors micro over strategy in short time, but the situation reverses over long time. Hereby, the preferred game length is decided by the skills the of the player, the length being such that it utilizes the persons competencies the best.

Creeping before, and only THEN attacking, would be theoretically the same as attacking right from the start. This situation is made different, if one of the persons has superior ability to benefit from creeping. I don't know why it is so, but I seem to have that skill. So I choose tactic according to my skills and avoid tactics that are bad for me.

There is a maximum time that one can use spending time with creeps; that's until everything is creeped. So my strategy isn't invalid by making the games take infinite time.

Again the original critique of the "rush=win!" claim stems from my observed tendency of the people to disregard other people's action. This is evident from seeing their posts on forums that discuss strategy. Good 60-70% of posts are in the form of, "make footmen and win, it's gg". These tactics disregard the possibility for the other player(s) to have a) anticipated the attack, and b) their capability to react. So any expressed tactics of this format make an unfair comparison.

But one point: these do work at the extreme noob level. Just taking a tactic that is not overly silly will beat the people who indeed do take overly silly tactics. Like that.. don't make mass wisps, make other units instead, like archer.

I think I understand more of the people's situation now and I can understand some people's preference to rush. If someone has an abundance of skill in the areas that the rush requires, he would have to utilize that skill in order to maximize the outcome. So the "rush=win!" exclamation can be seen as the equivalent, "my skills are best used in a rush. Not rushing would decrease my contribution to the game". But I still think that this approach still oversimplifies the person's contribution in the team; I think it's a communication that disregards other people's talents from the start. It is some freaking TJ trait! ETJ?

Maybe I have a bias against not being nazied and getting "facts" shoved down my throat. I still like the idea that we are not entitled to our own facts.

I do see myself as exploiting a hole in the rush strategy. One thing is also that people have preference for 10-15 minute games and prepare inadequately for 10 more minutes. My games take considerably longer on average, 20-25 minutes. I can understand that someone enjoys short games more than wins. I don't consider this exactly a flaw.

Just as a reminder, the bias was not about the preferred strategy, but about thinking that one's strategy produces wins over others, whereas the strategy is so usual as to actually produce 50:50 wins unless if it is refined. I think that people's skills, preferences and experiences bias them to think that what they experience themselves is as good as a law that is encoded within the game. I have also seen this bias expressed in many communications, so that people think to hold a superior strategy that will produce good outcomes, whereas other people hold similar strategies. Self-serving bias?

this 20% win ratio by the creeping tactic may have been your experience. I cannot verify it with my experiences.

 

[ptgatsby]

Creeping before, and only THEN attacking, would be theoretically the same as attacking right from the start. This situation is made different, if one of the persons has superior ability to benefit from creeping. I don't know why it is so, but I seem to have that skill. So I choose tactic according to my skills and avoid tactics that are bad for me.

I'm not sure I agree. Consider that the two strategies may not have the same dynamics.

Your strategy would work better at the macro level - a systemic growth with the requirement not to have a fall-through. It's a low deviation strategy.

Their strategy would work better at the micro level - a single point of failure allows them an overwhelming advantage.

(If you play magic, the analogy I use is Blue control vs threat management, where if blue can counter any threat, they are virtually untouchable, whereas a single threat can destroy them).

This leads me to believe that the strategies are asymetrical, in which one plays differently and has a different risk profile than the other.

But one point: these do work at the extreme noob level. Just taking a tactic that is not overly silly will beat the people who indeed do take overly silly tactics. Like that.. don't make mass wisps, make other units instead, like archer.

I think this emphasises my point above - your strategy is good... if you are good at it. However, the skill level and risk level doesn't scale the same. One mistake by you against a rusher can end the game - if you are good enough to prevent breakthrough, etc and still out resource/level/etc the enemy the majority of the times, then your strategy would be dominant (but this would leave the optimal strategy with those without your skillset to rush... and yes, I agree, that would be likely )

The inferior player has nothing to lose by rushing. The skilled player loses very little, if they are playing someone of equal skill... but that assumes that you could play a rush as skillfully as the creep approach, which is unlikely.

But I still think that this approach still oversimplifies the person's contribution in the team; I think it's a communication that disregards other people's talents from the start. It is some freaking TJ trait! ETJ?

I think this is natural. You also have a preference, one that I don't think many people can emulate. So it seems there are multiple factors. In the abscence of a developped strategy, rushing is optimal because it takes the least amount of skill and is the most likely to cause the "one error = lose" threat. However, and as a result, those that are unable to deal with the "one error" problem also rush as a counter because they don't have your skillset.

This is probably a relative skill level problem and at the pro level the rush is probably done to prevent your creep tatic (presuming the skill level there is high enough to avoid the one-error problem to start with) as much as anything else.

Just as a reminder, the bias was not about the preferred strategy, but about thinking that one's strategy produces wins over others, whereas the strategy is so usual as to actually produce 50:50 wins unless if it is refined.

True, but the question is if they are truly biased or measuring their own abilities. I think the bias here would be more conditioning - you get punished for not rushing, therefore training you to rush in the next game. Every "error" that produces a loss for you against a rusher encourages that mindset.

 

[Ghost of the dead horse]

This leads me to believe that the strategies are asymmetrical, in which one plays differently and has a different risk profile than the other.

So it would seem. (edit: not perhaps in the military sense asymmetrical. I a not so familiarized to the term in that sense.)

One mistake by you against a rusher can end the game - if you are good enough to prevent breakthrough, etc and still out resource/level/etc the enemy the majority of the times, then your strategy would be dominant (but this would leave the optimal strategy with those without your skillset to rush... and yes, I agree, that would be likely )

Yeah, I think that there's something to be learned from what people tend to do.

This is probably a relative skill level problem and at the pro level the rush is probably done to prevent your creep tatic (presuming the skill level there is high enough to avoid the one-error problem to start with) as much as anything else.

The rush utilizes the advantage in asymmetrical information. The rush is performed at a time known by the attacker, unknown to the other. So the attacker can optimize his things. Sometimes it becomes a decision problem for me, when I know that I can't be forced to be in my base, because I lose creeping, but if the other player pushes on me too hard, I can't leave it. This leaves the possibility for the other person to maintain the threat on me the most time, whereas he creeps more time than me, and steadily advances. Some people do this to me and I admire their skill.

True, but the question is if they are truly biased or measuring their own abilities. I think the bias here would be more conditioning - you get punished for not rushing, therefore training you to rush in the next game. Every "error" that produces a loss for you against a rusher encourages that mindset.

There it is, that might be it. I used to play rush in another game where I learned it to work better. I think there was a turning point in my skill development when I once played with malfunctioning system and learned not to trust the ability to execute fine movements. This conditioning stayed after I fixed my hardware.