Enter into the ∙∙∙{ Mind }∙∙∙ of a Pro - A guide on the psychological aspects behind every pro mind

[TaFoKiNtS]

.Very false dichotomy here. Usually hard reads put you out of position. At best if you miss, things revert to neutral, at worst the tables are turned on you. For example: Say your opponent has been doing the same thing with his approaches all match. You theorise correctly that the thing that would beat it would be wavedash back > attack. A hard read would be to wavedash back forward smash, something that would be a deep committal on your part. If this didn't work, then your opponent would likely be in a position to punish you. Conversely, a soft read would be to wavedash back > safe option. In Marth's case (as I am a Marth main if you couldn't tell) would likely be a dtilt. If the dtilt lands you can often convert it into a grab, and if nothing else you can press forwards and take a bit more stage. If this didn't work, you would likely not be punished afterwards seeing as it is safe. So yeah, a forward smash would have been nice, but you risk opening yourself up for something just as bad or worse. If things go well with the safe option, you can actually convert it into something quite good (like a grab) that actually isn't much worse and in some cases could even be better, all without taking as much of a gamble.

I would beg to disagree and it seems to me that we have very different opinions on what a hard read can be. What you present is an all-in situation that will make it seem like I will die if I'm incorrect. There are hard read situations that are very different and what I'm stating is that there are situations in which risk/reward can shift into going for a harder read rather than playing purely safe. (I do think we are confusing terms here and talking over each other in semantics).

I would give the example of Falcon dittos on a stage with platform. Let's say you have the chain grab on your opponent who is at 40-50% (basically a percentage in which uthrow --> uair is not really feasible). I can try to aim for a regrab scenario which would give me ~10% or aim for a stomp --> knee that gives me 40% and a potential kill. The stomp read is more of a dedicated read, but if I'm wrong, I still have control of the situation (perhaps a little less). If I go for another grab, this could lead to a damage cap of them getting on a platform or using some SDI/DI tactic to get them out. In this case, I think it's better to go for a stomp.

I never said reads were bad. I never said prediction was pointless. I said that reaction > prediction. Sometimes I do go for reads. Sometimes I can feel that, at a particular time, that I need to make a push, but this is not something that is common, and is still likely sub-optimal. In general, making several small victories is greater than landing one big one.

And yes, I agree with this assessment for the most part, which is why I said I agree with your overall premise, but I also wanted to add an * because there are situations (not-normative) that people should be looking for a read. I'm not saying to always go for reads/predictions

First, I'm not even going to speculate about why Hbox can beat x-players, but I doubt it's because of your claim. I think Hbox is actually really, really talented despite how much people loathe Jiggs and his playstyle. I actually think Hbox plays very safe and methodical, so I really don't see your point there.

Hbox has a safe methodical approach, but his edge comes from getting reads off of very predictable movement. Jiggs can't inherently cover all of a faster character's options, so some prediction is required, but i digress

Also, people don't seem to understand the idea that safe /= predictable, or that predictable == bad.
Like I have stated before, there are tons of scenarios in which it doesn't matter if you know something is coming or not, you cannot avoid it.

From my understanding, you're trying to quantify safe as minimizing risk, which implies some optimal moveset/movement, so I would disagree based on this premise that safe =/= predictable. Yes, there are unavoidable situations, but if that were every case, this game would have been solved. There are many mini-skirmishes that occur in a given game that require you to make a choice that may not cover everything as much as you want.

The whole point of my mentality, and really the new school Marth mentality, is to create these situations as much as possible. If a strategy is solid enough, it doesn't matter if your opponent is aware of it or not. Honestly people should re-evaluate why they do the things they do and try to seek these setups for their characters as well. Even if you are heavy on predictions, you want to remove the ability for your opponent to save himself. Also, playing safe does not mean to repeat the same thing over and over just because it is "safe". Most characters don't have many pokes, or great approaches anyways, so you KNOW what they will likely do in the first place. That doesn't mean that they won't still land it. Playing safe just removes the bad, over committing options that, in which case if I was wrong, (which in neutral it's BAD to guess wrong) one would be punished hard for it.

It's much easier to win out in neutral against someone who is risky than someone who is safe. If someone is risky then you just have to be patient. Sometimes they might catch you off guard, but more than likely they will give you all the things you need on their own.

To the new "Marth" meta, yes I would agree and it is the basis to any game at the root core. I'm not disagreeing with this base premise of your definition of "playing safe" and avoiding fruitless risk as I've stated before.

I think we have two different ideas of what "safe" and "risky" are. I'm not talking about randomly Falcon Punching or F-smash spamming repeatedly. I would also say that a person that is slightly risky is harder to work around than a person that plays to what I perceive to be the "safe" option every time.

 

[Bones0]

I like this discussion.

It was a good read.

 

[shadrach kabango]

these arguments are almost always so stupid because the answer is almost always the same:

combine the two. it's not predictions/reads or reaction, it's both. they are one and the same. if u dont understand this ur just not good enough imo (although many people understand this intuitively and just aren't so mercurially inclined to state it thru language)

 

[Ilove]

It doesn't matter if you are being read or not with a lot of player vs character strategies. If you are on a platform, and I am below you as Marth, I'm pretty sure that you will read that I will either uair or utilt you, but there isn't much you can do with that data because you are put into an incredibly bad position. Likewise, if I upthrow > juggle, most of the time the opponent knows I will attempt to juggle them, but that doesn't make their position any better or my position any worse. In a juggle scenario it's quite severely not in their favor, as their options are not only limited but finite. If you use your second jump and haven't secured safety then you just made things far worse for yourself. If you are playing to maximise every exchange you make while not relying on hard reads, you likely won't land as many big victories at any point of the match, but you will win small battles consistently throughout the match that will result in a win.

Prediction isn't just knowing what someone's going to do; it optimally is knowing exactly what they're thinking and executing in conjuncting with those thoughts. This means what move will be used, the timing, and the spacing. Though some situations are unwinnable once you're stuck in them (at least in theory), knowing exactly what you're opponent is doing in these terrible situations will allow you to see a hole if there is one, which wouldn't be seen if reads are not being implemented in disadvantageous situations. The uptilt and (especially) uair are certainly punishable if they are being done on someone who's standing up if someone can fully predict the timing/spacing.

Prediction on top of reactions is what allows these small victories at any decent level of play; pure reactions will only allow you to beat ******** players that over commit to uncertainties. If you over-simplify the game, you'll never fully understand it.

Prediction is as hard as it is made out to be, actually. Even people who are amazing at hard reads flub them all the time, and they often hurt as much as they help. Also, how is acting after your opponent counter-able? When you predict, you are being proactive and act first. When you act, you commit to something. If an opponent is capable of reacting to something, it will generally beat out a prediction because your move cannot be undone/changed at that point. In regards to people "baiting", just because you play reactive doesn't mean you just fall for baits all the time. This is where playing very safe and patient comes into play. Baiting is also something that is overhyped, because you are essentially relying on a strategy that depends on your opponent making a mistake, and often times going for a bait puts you out of position or removes options for you in order to perform it.

1. Clearly you don't understand prediction if you think reactions can completely replace them, so I don't think you have any knowledge based upon to come to such a definitive conclusion as to the difficulty of prediction. We have a neurological system that specifically allows us to empathize with the basis of the ideas behind other's actions; being in touch with this system can make prediction, at least of impulsical things, relatively easy and certain (extremely easy against players that don't adapt, which requires prediction in a different way). If people in significantly more advanced (not in complexity but in level of play) games like boxing didn't optimize this system at least to a considerable extent, they would get absolutely destroyed. Not everything is reactable, and if you sit in shield or try to play too safe you will inevitably be put in a position where you're forced to do something and then your reactions are worthless, assuming you're playing against a mentally/mechanically competent player.

2. When you predict, you aren't being proactive directly; prediction is the idea behind knowing what an opponent is going to do (not the act of punishing; that's something that could come from the idea). If an opponent is capable of reacting to something and is ONLY reacting, it means they will never do things until an opponent does something that the reacting person sees as punishable. There is never a situation where you can just avoid all confrontation; if you're waiting to react to something and an opponent runs right in front of you, what are you going to do? A predicting opponent could read what you're going to do and punish (even if your action is just a wavedash, run away, or shield; if you're waiting for an opponent you're giving him/her the upper hand in controlling the situation, and these things are punishable at certain times too). I know you acknowledged that prediction is necessary in the neutral game, but I don't think you realize just how strong it is at high levels of play; prediction becomes the primary mental idea virtually 100% of the time at neutral in high levels of play.

Playing patiently and safe is definitely important; but you can never be safe if you don't know what your opponent is going to do.

Actually, this is also untrue. Go take a peak at Kadano's Marth thread. There are a lot of situations in which are perfectly reasonable to react out of that people aren't really aware of. That's the whole point of eliminating options in the first place, so that your opponent is left with so few real options that it doesn't matter what they pick because it is possible to cover them. This doesn't mean that the reactive player messes up, we are all human, but when you rely on reaction instead of a guessing game, you will find yourself playing much more consistently.

That thread didn't really apply to your point at all (unless you're talking about true combos as reactable situations, which I was not; the only real thing that I saw in there that kinda applies to your point is the platform teching covered by uair, but either way I did say very few situations) and also the illusion section is horribly incomplete; there are too many timings/spacings and other recovery options that aren't being accounted for.

You can't limit a player's options unless you win the neutral game, which against requires prediction generally. You can't realistically know everything an opponent is going to do, but you can pick up on habits using prediction to force situations where it is more likely for you to get a punish. Predictions enhance reactions; both are very important. Prediction, when used correctly, doesn't equate to over committing, it merely allows greater chances to punish strongly both preemptively (if the read is certain, and if not no over-commitment should be in play) and reactively.

 

[MookieRah]

I like how I never once said that prediction isn't important, or that you could realistically cover every possible option of every possible encounter with reaction. I never spent much time in this debate talking about the fact that you can't cover everything with reaction (but I did address it), because it is blatantly obvious that one cannot react to everything ever. Hell, my opening post in this thread mentioned that there are only so many things a player is capable of keeping track of.

For those who don't understand my position, my point has been this:

A: Prediction is overrated, as if you rely on it too heavily it leads to inconsistency. That doesn't mean it isn't important.
B: Heavily committing to something on prediction alone is generally a bad practice, aka Don't get greedy.
C: Relying on reaction is better than prediction, when it is applicable.
D: Creating and seeking out situations that eliminate options from your opponent is one of the most important things in smash.

I felt the need to post on this purely because people were pretty much only discussing prediction, but nobody mentioned reaction, or even the idea to try to create situations where, by position alone, you could greatly limit your opponents options. I'd say those two things are very important, and quite clearly are undervalued in this community, and is something that Pro players are aware of (even if they can't really articulate it well).

I'm getting quite tired of discussing this now, as few seem to actually understand where I'm coming from. Instead of listening to my whole argument, most people are jumping on the "defend prediction" band wagon. This is also not a reaction vs prediction thread, this is a thread on someone's essay regarding high level play. At this point I think we've gone far enough with this discussion here as it is now taking over the thread and is disrespectful to the OP (yes, I realise I began this debate and am the biggest offender here, my apologies to the OP).

1. Clearly you don't understand prediction if you think reactions can completely replace them, so I don't think you have any knowledge based upon to come to such a definitive conclusion as to the difficulty of prediction.

This is insulting. I actually have made several posts before you even joined the smashboards on both prediction and manipulating your opponent.

Observation: The Key to Higher Level Smash Play: http://smashboards.com/threads/observation-the-key-to-higher-level-smash-play.103017/#post-2351003

Playing to Learn - http://smashboards.com/threads/playing-to-learn.176650/

Both espouse ideas that were very rarely talked about on the boards at that time, and mostly based on prediction-based things. Now people talk about the mental aspects of the game all the time, which is great. I honestly think that's in part why the bar has risen so much.

 

[TaFoKiNtS]

Have you ever thought about why people have repeatedly misinterpreted you and Umbreon? It may not be the audience that is getting it wrong.

 

[MookieRah]

Perhaps I could have been a bit more clear that I don't believe prediction is a bad thing, although I can quote a few times in this thread where I make note that it is not. People have a negative attitude towards reacting to things, and it is often misunderstood.

 

[Deleted member]

Have you ever thought about why people have repeatedly misinterpreted you and Umbreon? It may not be the audience that is getting it wrong.

I could not be more clear. Prediction in smash is bad and you should not use it. I'm being quite literal.

Prediction is an educated guess of what your opponent will do that you choose to use because you don't have the ability to cover the opponent's options. However, if you are wrong, this is bad news for you. It is generally better to stick to strategies that always work. In the event that you do not have that option, you can choose to go for something opportunistic if and only if the risk to you is functionally still zero. If at any point your prediction could actually have some adverse consequence should you be wrong, you should simply not take that risk. If you try to 50/50 your opponents on a hunch, 50% of the time they'll still beat you. Turning your matches into coinflips is not where you want to be in this game. And no, "predict better" is not the answer, if anything it's highly circular and frankly somewhat narcissistic. It is better for you to forego the risk and to re-establish your positioning so that your opponent is unable to put you into a disadvantageous position.

Yes, sometimes you will "get them" more, but sometimes you will lose out on it as well. However, it is a net positive to not even have the chance of putting yourself in a bad position to begin with. High risk / high reward scenarios are not good in a game where you must win a large number of rounds but can only lose twice. Simply do not do it.

Frankly I thought you'd be able to appreciate this perspective. A big part of why Sheik vs Fox feels so bad is because she has nothing reliable on fox from 0-40. You should understand first-hand how crappy it is to have to take risky guesses.

 

[MookieRah]

The problem in this thread now Mow is that the definition of prediction has become so broad to encompass many other things that I don't feel it would apply to. Hell, one guy was attributing that it requires prediction to space (basically I assume that he meant that you had to take into account such things as falling speeds, air speed, or ground speed and what your opponent can do slightly before and onward prior to making the input of your attack). At the very least, it's clear that the definition that we are using for prediction does not match up with most of the opposition's viewpoint. Defining it for the sake of this discussion didn't work either, as it was more or less ignored.

I defined it here, based more on how the thread was discussing it prior to most of the debate:

Spoiler

Prediction, as it's being used in this thread, refers mostly to reads (usually hard reads). It's basically referring to a player vs player style, one which adapts to your current opponent and attempts to manipulate them through means of baits, conditioning, and all kinds of mental tricks.

Reaction is more of a player vs character strategy. It's based on knowledge of what you can do and what your opponent can do in that given matchup. It's about abusing everything your character can do that gives you an advantage over your opponent's character. This kind of style does not rely on getting to know your opponent as much, as the strategies and techniques you employ would work regardless of their skill level or awareness of your strategy.

I think prediction is fine in the sense that you are simply identifying patterns in your opponent's play, but it shouldn't dictate what you do in a match more than giving you extra insight.

And no, "predict better" is not the answer, if anything it's highly circular and frankly somewhat narcissistic. It is better for you to forego the risk and to re-establish your positioning so that your opponent is unable to put you into a disadvantageous position.

You know I was trying to find a way to explain this, but couldn't quite find an adequate way to get my point across. I'm glad you nailed it for what it was, which is narcism. This viewpoint also fails to take into account that people with equally good prediction skills would, in turn, neutralise the effectiveness of prediction, but that seemed to be absent in the "get better with prediction" mindset.

 

[shadrach kabango]

lol @ umbreon speaking so authoritatively without establishing credentials to begin with

what's your tag? what's your tournament history? what significant players have you beaten?

i personally skip all your posts and will probably add you to the ignore list soon, but you dodged this question in the other thread, too, so consider this the call out.

 

[TaFoKiNtS]

The ideas you present Umbreon are based on the premise that a failed prediction = lose, whereas you can set up mixup situations in which you win less vs win more depending on predicting correctly. The reactionary measures will set up the scenario where you win either way, but I feel a prediction is the difference between a Fox trying to tirelessly shield pressure a Peach's shield, having to execute 20 maneuvers vs. a Fox simply grabbing, knowing that he's conditioned the Peach to shield.

^maybe not the best example here, but all prediction/read scenarios don't have to be leveraged in a way that if you're right, you get small punishes and when you're wrong, you flat out just get dominated, which is what you seem to be fixated on.

Game theory 101

Let's say you can take add a punish tilt for a guaranteed 10%, but they recover back to neutral, but you can go for a prediction in which

1.) 50% of the time you rack up 50% damage, they go back to neutral
2.) 50% of the time you rack up no damage, they go back to neutral

It seems you are predicated to not even bother with the 50/50 scenario here, even though a large sample size warrants going for the risk/reward read here. I'm not saying this is always going to happen, but my caveat was to say that you can set yourself based on reaction for these scenarios in which it isn't always wise to play reactionary and the reward for taking a risk can pay dividends.

In the Sheik Fox scenario, I feel like the game digresses into me being in quick sand in that unless I take some sort of risk in getting hits in, they will gradually overwhelm me and rack up damage. I can't simply play risk free against a character that outmaneuvers me and has the option of outpriortizing my aerials and putting me into a corner. I make it work because I find 50/50 tech chase grab scenarios when Fox is in shield, so it's a give and take =/

 

[Rarik]

Your analysis either doesn't take into account stage control, or seems to assume that it doesn't change at all in any of the options which I feel is incredibly unlikely in this game. Resultant stage control from these options makes or break which of the options it's best to go for. If the guaranteed option gets me 10% and allows me to gain stage control or otherwise puts me in a better position, and the "bad" side of the 50/50 doesn't, then I would take the guaranteed option every time. Better position results in either better options for me, or worse options for them. It could also let me escape a bad position by allowing me to contest/regain center stage, where otherwise I would stay in a bad position.

If they did give equal amounts of stage control, or no change at all, then I would agree with your statement that a large sample size does warrant going for risk/reward, however that seems very unlikely to me since the guaranteed options involves hitting them and the "bad" 50/50 doesn't.

 

[Bones0]

I could not be more clear. Prediction in smash is bad and you should not use it. I'm being quite literal.

Prediction is an educated guess of what your opponent will do that you choose to use because you don't have the ability to cover the opponent's options. However, if you are wrong, this is bad news for you. It is generally better to stick to strategies that always work. In the event that you do not have that option, you can choose to go for something opportunistic if and only if the risk to you is functionally still zero. If at any point your prediction could actually have some adverse consequence should you be wrong, you should simply not take that risk. If you try to 50/50 your opponents on a hunch, 50% of the time they'll still beat you. Turning your matches into coinflips is not where you want to be in this game. And no, "predict better" is not the answer, if anything it's highly circular and frankly somewhat narcissistic. It is better for you to forego the risk and to re-establish your positioning so that your opponent is unable to put you into a disadvantageous position.

Yes, sometimes you will "get them" more, but sometimes you will lose out on it as well. However, it is a net positive to not even have the chance of putting yourself in a bad position to begin with. High risk / high reward scenarios are not good in a game where you must win a large number of rounds but can only lose twice. Simply do not do it.

Frankly I thought you'd be able to appreciate this perspective. A big part of why Sheik vs Fox feels so bad is because she has nothing reliable on fox from 0-40. You should understand first-hand how crappy it is to have to take risky guesses.

What if you have a good read on your opponent and ARE able to predict his techrolls or whatever with >50% consistency?

What if a Puff throws someone off stage at 0% and have a 33% chance of guessing their recovery option and gimping them. Even if there is a 66% chance of the opponent getting back on with stage control, as long as Puff can regain stage control with a success rate >33%, that means it's worth predicting for the gimp.

 

[Deleted member]

What if you have a good read on your opponent and ARE able to predict his techrolls or whatever with >50% consistency?

What if a Puff throws someone off stage at 0% and have a 33% chance of guessing their recovery option and gimping them. Even if there is a 66% chance of the opponent getting back on with stage control, as long as Puff can regain stage control with a success rate >33%, that means it's worth predicting for the gimp.

This is a good idea, but it's flawed for a few reasons that I think you'll be able to appreciate.

The first reason is the marginal risk. Let's say that in your scenario, jigglypuff can regain control 100% of the time. Therefore, the 33% chance of a kill is effectively 0 risk. In this case, I would say that yes you should go for it. Now let's do the opposite extreme where it works 51% of the time at the margin that you suggested. The problem here is that in this model, there is no opportunity cost for alternative actions. At some point between it working 50% and 100% on paper, we hit an arbitrary threshold of risk where it becomes acceptable or unacceptable. However, choosing to forego that risk in favor of re-establishing position is generally a very strong thing to be doing at any point in the game, especially in your scenario where the jigglypuff player has already established a good point of position. And of course, the choice to re-establish position should fall as an equivalent to some risk threshold on our theoretical scale between 50% and 100%. Since the alternative of re-establishing position is a very strong choice, I would say that it is close enough to the 100% mark that any attempt at non-guaranteed strategies should be essentially risk-free. And really, that's what the best mix-ups are already; you're not really desiring a 50/50 when you're using positional advantage to force your opponent to choose, as you reasonably expect it to work out in your favor regardless of how the opponent reacts, so it's "nearly guaranteed" if that makes sense. You can basically use this idea to explain why things like frame traps and resets are so good. Even then, you really only want to use these types of strategies when nothing guaranteed is available.

The second reason is that your success rate will translate to your match win rate. So for example, if you win every exchange like that 66% of the time, you should average going something like 4-2 in a simple bracket. The problem with this is that you're playing to win the event with a record of X-1 at worst. If your bracket is 8 rounds, this is unacceptable. It is clearly the best option to go X-0 and to have a 100% win rate, which is going to be best supported by using strategies that work 100% of the time. This is slightly misleading because the game does not determine win quality, so you can win last stock in all of your sets but... that's not really the point. The point is that taking risks that you don't need to will cost you margin when our tournaments functionally disallow marginal loss if you want to advance.

Let me know what you think.

 

[Bones0]

It is clearly the best option to go X-0 and to have a 100% win rate, which is going to be best supported by using strategies that work 100% of the time.

No such strategy exists, or at least not any I've ever seen. Almost any strategy, whether it's clearly optimal or not, can be defeated by the opponent predicting you using that optimal strategy. If they can defeat your strategy by predicting it, then it's not a 100% success rate. If it still has like a 90% success rate (like Marth just waiting outside of someone's shield ready to do something on reaction), then yeah, keep abusing that every single time you're in that situation. If you're in a tech chase situation where the truly optimal course of action is to predict your opponent correctly, you have to weigh the risk and reward of guessing right vs. guessing wrong. This involves prediction, which is why I can't agree with your assessment that prediction is bad in Melee. Even if your prediction skills aren't much better than 50-50, that might still be better odds than your macro (minus prediction) vs. your opponents macro (plus prediction in low risk high reward scenarios).

 

[Grim Tuesday]

The problem in this thread now Mow is that the definition of prediction has become so broad to encompass many other things that I don't feel it would apply to. Hell, one guy was attributing that it requires prediction to space (basically I assume that he meant that you had to take into account such things as falling speeds, air speed, or ground speed and what your opponent can do slightly before and onward prior to making the input of your attack).

mookie you clearly don't understand the mental side of the game well enough to write multiple threads on it
if you don't try and predict the opponent as a component of your spacing game

and if you do, do you call it something other than prediction?

Prediction, as it's being used in this thread, refers mostly to reads (usually hard reads). It's basically referring to a player vs player style, one which adapts to your current opponent and attempts to manipulate them through means of baits, conditioning, and all kinds of mental tricks.

Reaction is more of a player vs character strategy. It's based on knowledge of what you can do and what your opponent can do in that given matchup. It's about abusing everything your character can do that gives you an advantage over your opponent's character. This kind of style does not rely on getting to know your opponent as much, as the strategies and techniques you employ would work regardless of their skill level or awareness of your strategy.

there is so much more to being in your opponent's head than 'mmhhm hard readzz mmkay'
what you expect your opponent to do should factor into ALL of your decisions to some extent
what you want your opponent to do next should factor into all of your decisions as well i.e. baiting and conditioning

and even with the hard reads thing dude
you keep talking about hard read options as if they work really well 1 in 10 times, but get you punished 9 out of 10 times. That's not a hard read, that's a guess.
That's like saying people shouldn't play reactive because timing shine out of shield is too hard and you'll **** up and get punished lol

If you go for a hard read and you're wrong, that doesn't necessarily mean you shouldn't go for it next time, it means you should actually get the read right next time lol

I think prediction is fine in the sense that you are simply identifying patterns in your opponent's play, but it shouldn't dictate what you do in a match more than giving you extra insight.

You can't separate prediction and reaction like this :/
Neither side gives you 'extra insight', I don't know how you can look at the game like that and actually quantify how much of each 'thing' you're using; as even the smallest decision involves both knowledge of what your opponent is thinking (or what you think they are thinking) and knowledge about the game's mechanics

 

[MookieRah]

Since people don't claim that my or Mow's position on this is clear, I'll do some more defining. I think people still can't distinguish some of the things being discussed because of the ambiguous nature of the word prediction.

mookie you clearly don't understand the mental side of the game well enough to write multiple threads on it
if you don't try and predict the opponent as a component of your spacing game

First, before I go any further, did you actually read any of the posts I wrote on the subject? They are in fact all about getting into your opponents head. I wrote a lot about the mental side of the game, stuff like establishing fear in your opponent, discussed baiting strategies, and even talked in a broad way so that people could understand the concept and construct their own strategies. These posts were also made before these types of concepts were discussed on the boards.

Now, moving on. There is a distinct difference between predicting something based on knowledge of the game versus knowledge of the opponent. One can easily make spacing decisions based on knowledge of the game mechanics as well as the character matchup. Lets call this, for purposes of this discussion, K prediction, or knowledge based prediction. This is usually not what people discuss when they bring up prediction in smash or other fighters. Most of the time prediction is discussed as "getting in your opponents head" as you said. Lets call this P prediction, or player prediction.

and if you do, do you call it something other than prediction?

It is "predicting" as any act of planning ahead falls under "prediction" but I want to make the clear distinction between K prediction and P prediction.

So now that we have a definition for these two similar but different entities, we can really discuss things.

you keep talking about hard read options as if they work really well 1 in 10 times, but get you punished 9 out of 10 times. That's not a hard read, that's a guess.

That is what a hard read is. It may be an educated guess based upon your interactions with a particular opponent, but it's still a guess. I'd also love for you to quote me where I gave numbers like "1 in 10" or get you punished "9 out of 10 times".

That's like saying people shouldn't play reactive because timing shine out of shield is too hard and you'll **** up and get punished lol

Is this really your retort to my arguments? Tech skill is not a factor in this discussion. Like I have mentioned, there are several cases in which one can learn to react to things on command, and they all fall within a fairly average reaction time.

Choosing not to do a particular option because you do not possess the proper technique to do it consistently has nothing to do with whether or not Prediction K is or is not better than Prediction P.

If you go for a hard read and you're wrong, that doesn't necessarily mean you shouldn't go for it next time, it means you should actually get the read right next time lol

Your opponent can adapt to what you are doing. I think it would be incredibly foolish to assume that you can do the same thing over and over and your opponent not "read" what you are doing.

You can't separate prediction and reaction like this :/

Yes, you can. Especially now that I have clearly defined the two separate things. It's now fairly simple. I can even give a clear example:

The scenario:
Player A, Marth, up throws Player B (can be pretty much any character aside from Pika/Pichu) on Battlefield from below a lower platform. The percentage is such that they will land on the platform. Player A notices that Player B has a tendency to roll away from you on platforms.

Prediction K - Input a short hop. Watch for tech, or lack thereof. If no tech, perform uair early. If a tech, perform uair at the appropriate time. This works because if you are positioned under the platform, uair can cover pretty much every option on reaction.

Prediction P - Acting on the prior knowledge of his opponent, Player A decides to cover the side away from him as Player A suspects that Player B will continue this pattern. Player A places himself to take full advantage, possibly even landing an fsmash tipper on reaction if Player B does indeed tech roll away.

I don't know how you can look at the game like that and actually quantify how much of each 'thing' you're using

You could quantify it with actions. In the above example it shows a clear difference, one that an outside person could tally. That said, you'd have to be very knowledgeable on a specific character (and that characters' matchup) to be able to quantify how often they go for a Prediction K or a Prediction P. That is not my point though, my point is that it is obvious to your own person if you opt to go for K or P. I advise to stick to K as much as possible is all.

as even the smallest decision involves both knowledge of what your opponent is thinking

My above example is an example in which a player's decision has no bearing on what happens next. As long as Player A is on point with tech skill, it is a given that Player B will be hit by an uair. It doesn't matter if the opponent misses a tech, techs in place, or rolls, the uair can cover all the options.

 

[Grim Tuesday]

There is a distinct difference between predicting something based on knowledge of the game versus knowledge of the opponent. One can easily make spacing decisions based on knowledge of the game mechanics as well as the character matchup. Lets call this, for purposes of this discussion, K prediction, or knowledge based prediction. This is usually not what people discuss when they bring up prediction in smash or other fighters. Most of the time prediction is discussed as "getting in your opponents head" as you said. Lets call this P prediction, or player prediction.

It is "predicting" as any act of planning ahead falls under "prediction" but I want to make the clear distinction between K prediction and P prediction.

No good player makes decisions purely based on game knowledge. There is no K prediction and P prediction, there is KP prediction and that's it. Sometimes you'll be using more game knowledge than opponent knowledge when it comes to making a decision, players probably lean towards one style more than the other as well, but they are always working in tandem.
If someone told me to choose my next option based entirely on game knowledge and ignore what I know about my opponent, I wouldn't be able to. It's too automatic and deeply ingrained in my decision making process to ever ignore it, and it'll always factor into my timing and minute spacing adjustments even if it doesn't affect the move that I choose itself.

That is what a hard read is. It may be an educated guess based upon your interactions with a particular opponent, but it's still a guess. I'd also love for you to quote me where I gave numbers like "1 in 10" or get you punished "9 out of 10 times".

I'll take educated guess, that's an apt term. Sometimes it is wise to go for hard reads.
A really clear example is G&W in Brawl; he legitimately needs to disrespect the opponent with hard reads to win because of how telegraphed all of his KO moves are. A G&W player who relies on always playing safe and choosing the best option based on game knowledge actually performs worse under tournament conditions, generally, than a G&W who throws out hard read smash attacks.

It's really clear in Brawl because usually in that game you DO want to play 100% safe, basing your decisions more on game knowledge than player knowledge (though it's still a factor, it always is) because of the stronger defensive game, slower pace, etc... In melee, though, every character could benefit from going for hard reads when the situation calls for it.

I don't need to quote you, by the way, because I never claimed you said those things.
I said you talk about hard reads as if they only work 1 in 10 times, and '1 in 10'/'9 in 10' are obviously hyperbolic figures of speech and not exact ratios that I've figured out lol

Is this really your retort to my arguments? Tech skill is not a factor in this discussion. Like I have mentioned, there are several cases in which one can learn to react to things on command, and they all fall within a fairly average reaction time.

Choosing not to do a particular option because you do not possess the proper technique to do it consistently has nothing to do with whether or not Prediction K is or is not better than Prediction P.

Perhaps not, but my example does demonstrate the absurdity of blaming the move just because someone used it poorly.
A hard read is an educated guess like you said, educated in the sense that you think your opponent is going to do whatever you're reading them to do, and you also think that even it's worth going for the risk even if they don't do it. Just because some people suck at weighing up situations like this, doesn't mean that hard reads are a bad option any more than sucking at tech skill makes shine OoS a bad option.

I would never suggest choosing a sub-optimal option for style or anything like that, which seems to be what you are interpreting 'hard read' as? Some ballsy thing that will work once in a blue moon and gets you insta-KO'd if you're wrong? In some situations, hard reads are the optimal or at least a viable option.

Your opponent can adapt to what you are doing. I think it would be incredibly foolish to assume that you can do the same thing over and over and your opponent not "read" what you are doing.

I didn't say do the same thing over and over, I said get the read right. That would, - more likely than not - involve doing something completely different.

I can even give a clear example:

The scenario:
Player A, Marth, up throws Player B (can be pretty much any character aside from Pika/Pichu) on Battlefield from below a lower platform. The percentage is such that they will land on the platform. Player A notices that Player B has a tendency to roll away from you on platforms.

Prediction K - Input a short hop. Watch for tech, or lack thereof. If no tech, perform uair early. If a tech, perform uair at the appropriate time. This works because if you are positioned under the platform, uair can cover pretty much every option on reaction.

Prediction P - Acting on the prior knowledge of his opponent, Player A decides to cover the side away from him as Player A suspects that Player B will continue this pattern. Player A places himself to take full advantage, possibly even landing an fsmash tipper on reaction if Player B does indeed tech roll away.

So why wouldn't you take your prior knowledge into account for Prediction K? Wouldn't a decision made with ALL available information be superior to a decision made with only some? Player A can aim to cover every option on reaction (even if it just means re-positioning himself, rather than a direct punish) including the option he expects - which will actually be incredibly easy to cover because you react to something even faster when you think it's coming.

---
btw if my post is too dry, just watch this instead

not because it demonstrates the importance of reading, as pretty much every single competitive melee video already has that in spades, but just because it's awesome.

 

[Deleted member]

No good player makes decisions purely based on game knowledge.

I like how this is just strictly false lmao

 

[Grim Tuesday]

pics or it didn't happen

even if a player ACTED purely based on game knowledge, they still had an expectation of what their opponent would do which they factored into their decision, it's pretty hard not to.

 

[Zumbach]

Part 2/2 Video unavailable