I guess this is more me being curious than anything else, as I rarely play with stacks much deeper than 150bbs, but I have the question on my mind and it won't go away without an answer!

For 100bbs I think reasonable preflop raise sizes look something like this:
Raise to 3 bbs, 3-bet to 10 bbs, 4-bet to 25 bbs, 5-bet all-in.

Starting from the same point, a preflop raise to 3 bbs, what does this look like for other stack sizes?

150 bbs?
200 bbs?
250 bbs?
Beyond?

Are there any rules of thumb for this? At the very least if I know what reasonable raise sizes look like I also know what unreasonable raise sizes look like, which seems to me at least marginally useful. I know, I know, we shouldn't look to generalize etc., etc., but if anyone is willing to entertain me I'd appreciate it.

I would say as long as the stacks are deep enough 3bets, 4bets,5bets,etc. sizing remain almost the same regardless whether u r 150bb deep or 300bb deep although u will have to increase the sizing a little as stacks get deeper. Experience will tell u the proper sizing as it also depends on ur opponents and the general table

I'm not sure what they should be, but as you get deeper than 150bb, there are obvious implications of 5betting the same size to 100bb.

Perhaps those implications would warrant some tweaking of our 2bet, 3bet,4bet, and 5bet sizing altogether.

We can definitely 4bet larger when deep without be forced to commit preflop.

The "standard" 4bet sizing of 20-something BB is, I think, designed to allow us to 4-bet-bluff. After all, if we 4-bet to 30BB, we're more or less pot-committed to call a shove. With deeper stacks this doesn't hold. For each specific stack size there would be issues of pot geometry that come into play, but I think that with infinite stacks, a standard 4-bet size would be 3 times the 3-bet size, i.e. 30BB, a standard 5-bet would be 90BB, and so on.

Of course, pre-flop raise sizes also depend on position, and should usually be larger OOP and smaller IP, so the above sizings should be adapted according to this. Furthermore, there are specific ways to exploit one's opponent (bet bigger for value and smaller as bluff vs fish, etc).

why we raise bigger oop and smaller in position?

i'd rather want to play smaller pots oop then bigger ones, since we going to loose them

^^^^ ah, this is a time-honored debate. |I think raising bigger OOP is generally better, and I'll explain my reasons, but I don't have air-tight theoretical justification. Furthermore, I think we can find villains and/or situations where we should raise bigger IP and smaller OOP. But I do think that usually, one should 3-bet larger OOP.

Here are my reasons: First of all, let's assume for simplicity, for now, that villain calls every OOP 3-bet we make (I'll remove this assumption later). So, for each hand where we try to decide whether we want to call or 3-bet, we just want to decide whether we want to play a big pot or a small pot. Now, if the hand was -EV, we should fold it anyway. So we should only play it if it's +EV, and if it's +EV, we'd want to make our profit bigger, and we do that by 3-betting. There are two problems with this analysis:
1. I ignored the fact that we put a BB (or SB) already in the pot, so a hand might be +EV when we have already put in the blind, but -EV otherwise. But these are the weak hands in our calling range, the ones we won't 3-bet anyway, so that's fine.
2. Issues of balance: above I discussed deciding whether to 3-bet a hand or not in a vacuum, but in fact ranges and balance need to be considered, and our goal should be to be the most +EV in general. For example, it's possible that our 3-bet range is "too well defined" so the fact we 3-bet gives villain some info, which might turn a +EV hand into a -EV hand. So, we need to be careful about constructing our 3-bet range such that this doesn't happen, and we can include in it some hands that are slightly -EV in a vacuum, just to balance it out. Similarly having a 3-bet range also manipulates our calling range. So, overall, we should construct a 3-betting range in a way that makes our bottom line as good as possible. But, in this 3-betting range, I think most or all of the hands in it will be +EV individually, and therefore we want to 3-bet them large, to make their profit multiplier higher.

So, overall, as you can see, our 3-bet range from the blinds in this case should be +EV in general, and we 3-bet to increase our "profit multiplier". Another way to see it is that our 3-bet range OOP is supposed to be pretty strong, and with a strong range we 3-bet larger for value. (Note that we don't have any 3-bet-bluffs per se, since villain always calls; we just have hands we add to our 3-bet range for balance reasons, but these shouldn't change the fact that our 3-bet range as a whole is significantly +EV, otherwise we'd just get rid of that range and never 3-bet).

Now, what happens when we remove the assumption that villain always calls our OOP 3-bets? Now things get complicated, since he can start folding more when we 3-bet too large, and he also can start 4-betting. I think that real-world villains actually call 3-bets with a large enough frequency, and their flatting range is inelastic enough, that the arguments from before still work. In other words, our range should still be mostly value, and thus we still want to make the hand as big as possible, and we're not too sad if villain decides to fold to our 3-bet: all we want to do is to build a (probably merged) 3-bet-bluff range to exploit his fold-to-3bet frequency. This is roughly where my theoretical justifications end. I feel that our range here would be strong enough and in it we have enough card advantage that we should 3-bet large for value, but I don't have a good enough model to prove that is the case.

Also note that when we 3-bet bigger, we give villain worse implied odds, which achieves one of two things: either he calls with too many hands, making his range weaker and giving us card advantage that gives us more value, or he calls with a smaller range, which define his range better and helps compensate for his positional advantage (If I'm not wrong, positional advantage is less useful with smaller ranges).

But you are right that a major reason to 3-bet big OOP when stacks are 100BB is to make the SPR smaller, which improves the situation of the OOP player, and that this reason does not hold anymore when stacks are deeper. However, as I wrote above, I think the other reasons still hold.

Yeah we can't 3b as much oop so we'll have a stronger range, he can continue with a worse range, and there's more value to winning the pot now/keeping SPR smaller when oop.

noone over3bets here for value?

like making it 20 BBs when being 200 BB deep

i mean, that way we cut there implieds on us, if we only 3bet to 9-10 BB we are lost

wbuster, I struggle to see the connection between your last post and the overall discussion. Can you give more details and context on what it is that you're asking?

the hole point of 3betting oop is that we want to generate value with our good hands (beside throwing in some random 3betbluffs in our range)

but the thing is, if the spr is very high on the flop, even after we 3betted, we just put us in a spot where we generated value for our villian, since the spr is so high, he can pull out the crazy "i know u got probably only Ahi on this board"-bets, where on the other hand we could just potshove even with "only" Ahi if the spr is low

so if we only 3bet good hands oop(at least the most hands are "good") our villian has so much information, that he can raise us every flop and we are just playing the guessing game even if we have lets say AA

he knows that we have good hands in our range, but he knows too, that most of this hands didnt connect on most boards, hell he could raise every Axx flop if he knew we only 3bet PPs, broadways and broadwayaces, since in 2/3 of the cases we dont like that ace

just random thoughts^^

Your view is not logically consistent, I think.

There are two options:
1. either we carefully construct a 3-betting range and a postflop strategy that will be +EV.
2. Or we conclude that we cannot do such a thing, and decide to never 3bet preflop (even with AA).

You try to argue that #2 is the case, since you argue that our big 3-bets just make us lose bigger pots. But in this case, we just shouldn't be 3-betting at all, it wouldn't be any good to 3-bet small.

I personally have a feeling that #1 is correct against most (or all) villains, and in this case our 3-betting strategy is +EV, and I believe that in this case, our 3-bet sizing should be on the large side, for the reasons I stated in my previous posts.

i dont say its wrong to 3bet, i just think that 3betting our valuerange oop is kinda wired when ultradeep

i just think that playing deep is a trapgame, and also a overbetgame, you would be suprised with all sorts of trash they call overbets otf and turn (and some ******s on the river), so why not just start preflop with the overbetting?

if a fish thinks he can outplay you for additional 6BB he may think he can do it too for 17BB

the counterstrategy to flatting 3bets wide is, i agree, making the 3bet bigger, but not from 9BB to 12BB, its best to make it so big, that the villian wants to pay to see those 3 cards

also, if villian wants to 4bet, the 4bet has either to be only a min3bet, where we could setmine with QQ, or has to be really big, where we could jam over the top with AA

so either way i like

I'm sorry, I don't know games where a pre-flop raise from 3BB to 20BB makes sense, but maybe I don't play in your games. Regardless, since this is theory forum, I suspect that people are more interested in arguments that are of theoretical nature, rather than "this is how it works in the games I play".

if a fish calls 200BB shoves with JJ why wouldnt i ship KK+ vs him?

in theory of course
theory without the possibility to test it is no theory, its a belive