Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This author doesn't have much experience with competitive gaming and it's quite obvious. This is only a strategy he devised to play against his friends and relatives, most probably. He assumes a lot of player behavior, and what other explanation could there be? His bathtub example is very bad because there ain't no players, and when there's players the meta absolutely depends upon who you are playing against. there's not such a thing as an optimal universal strategy at all, in fact it makes a player predictable and low-skilled in most competitive games.

Systems thinking without game philosophy understanding is very shallow.

To illustrate my point of the strategy being entirely reliant upon opponents (meta): in cs:go, most unexperienced players will go through a hot spot (that's probably being targeted by pre-positioned opponents) without jumping, making themselves easy targets. this is the behavior you will find in low ranks. eventually, some of the players will learn that jumping may be a good tactic in these situations instead, to make it harder for opponents to hit headshot, and behavior becomes very common in mid-ranks. Eventually, though, mid-rank players will start to notice this tendency to jumping, and will position their crosshairs looking for a jumping headshot - they will progress in rank by doing so. At high-ranks, however, a lot of times it is expected that your opponent will be waiting for a jumping cross, and therefore the low-rank behavior would be the best one: don't jump. In conclusion, it's impossible to point out the best possible behavior without knowing who you're playing against and their skill in the game.



I completely agree with everything you said, however I'd like to comment on the last thing you said - Just because the best possible strategy changes depending on the skill of your opponent(s) does not mean that there can't be a strategy that cannot be exploited by your opponent. I think the mistake many people make is that they tend to look at games with a "chess mindset" (there is a single move that is the perfect response to my opponents move) when games with hidden information (such as cs:go) will require a "poker mindset" - a strategy isn't given by a single move but by a probability distribution of actions. In your example, there is an "optimal" way to play this scenario in which you aim at jumping height p% of the time and at normal walking height (1-p)% of the time where the value of p depends on many different factors such as you hitting the headshot if you aim correctly, you hitting it anyways if you don't aim correctly etc.

This is not the strategy that will win the the most (if you know for sure that your opponent will jump every time, of course you would never aim at walking height), but it is one that cannot be exploited by your opponent deviating from whatever their optimal strategy is.

Of course, improving your skill enough to see these changes in behavior in your opponents takes a long time, which is why we naturally adapt the more exploitative strategies that you mentioned.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: