I was asked via message about how to choose an open raise size at different stages of multi-table tournaments. My response was long, so I decided to post it here in case anybody else might find it helpful. Generally speaking, open sizing should be determined by two factors: stack depth and range width. From there, it is possible to get pretty scientific about determining an optimal sizing, but I prefer to think about it logically instead of mathematically. The highest level of precision is only needed against elite players, and there are none of those on Replay. Position can be another factor, but particularly in shallow-stack tournaments like on Replay, I like to use the same size from all positions.
Stack depth and range affect one another, but I will try to talk about each of them separately. Replay tournaments are essentially all turbos (6-8 minute blind levels), so stack sizes are usually quite shallow (<40 bbs). It is good to raise larger in the early stages of tournaments because players like to call too much and psychologically the costs of seeing a flop seems lower (calling a bet of 90 or even 250 chips seems like nothing). But, while players on Replay call too much in the early stages (and in ring games), they actually play too tight-passive and fit-or-fold once the stack depths get shallower.
What this means is that they will have already decided what range of hands they will call a raise with, and their decision will not be flexible based on the size of that raise. For instance, if you have 9,000 chips (100/200 blinds) and you open to 400, players will call/raise with the same range of hands as if you had opened to 600. They also will frequently fold to a c-bet when they miss the flop. So, I exploit this by opening a wider range of hands to a standard size of 2x (particularly from late position). I can often pick up the blinds and antes uncontested (which is especially great when stacks are shallow), and I can often win with a flop c-bet even when I miss. An alternative approach would be to open larger, but only with extremely strong hands, so players who call too much will lose large percentages of their stack against your strong range, but this approach has disadvantages of: not being able to play as many hands (losing out on many blinds and antes), making your range pretty obvious, getting stacked when your opponent flops well, and less postflop maneuverability.
Another reason why opening smaller is better is because of short-stacked players. If you open to 2x it is pretty easy to fold versus a shove (unless you have a big hand) and you don’t have to consider awkward situations where you may be pot committed. For instance if you open to 400 (8,600 behind) and someone after you shoves 2,200, you have to call 1,800 to win 2,900 (assuming no antes), while if you opened to 600, you need to call 1,600 to win 3,100. The amount of equity you need to break-even (not that making break-even all-in calls is a good strategy) goes from 39% to 35%, which makes your decision much more difficult with a wider range of hands. When I assigned my opponent a shoving range and calculated what % of starting hands have the right price to call in either spot (39% equity or 35% equity), the number of hands that should call increased from 13% to 23%, meaning that you would be facing a very dicey decision with a large number of starting hands. By opening smaller, you make your calling decision easier.
Which brings me to hand ranges. Especially while learning it is fine to only open extremely strong hands, particularly since many of your opponents here will call too much. But this makes it tricky because if you only open say QQ+/AK, as many players here do, then a 2x size is too small to generate enough value and to avoid multi-way pots given your extremely strong range. You will not be playing very many pots and when you do open, your strength will be pretty obvious. On the other hand, if you open 80% of all starting hands (which is only advisable in short-handed situations), then you want to open to 2x because you don’t want to build large pots with terrible hands and lose big when you have to fold against 3-bets. Exact ranges vary based on the situation (number of opponents, tendency to call, stack depth). But if you use the 2-2.25x approach, you can open a ton of hands (Ax, suited connectors, pocket pairs, suited broadways), not just premiums. Your range will be strong enough if somebody starts pushing back and not so loose that you look like a complete maniac. Later in MTTs, I also tend to weight my range towards top-pair and blocker type hands, even bad ones like A7 or KT rather than some suited connectors that I would favor more in deeper stack played (like 96s or 54s) because there is less room to maneuver and less fold-equity postflop.
In summary, smaller sizing is less risky (exposing less of your stack when you open a range that cannot call many shoves), it gives you more maneuverability postflop, it can still get folds versus tight-passive opponents who predominate late in tournaments after the bingo fish have crashed out, and it enables you to get involved more often with a wider range of hands. If you struggle postflop, then maybe a larger sizing and more straightforward approach could help avoid postflop mistakes, but to build an effective strategy you want to set yourself up to use your postflop skill to your advantage.
And one final note, if you have under 20 big blinds you should usually just go all in. There are calculators that will tell you whether or not shoving is correct or not in a particular situations, but I think on Replay (particularly because the stacks get shallow so quickly) it can be advisable to be a little trickier in some spots. For instance, my advantage over most players is in my decision-making after the flop, so while shoving 20 big blinds with KTs is possibly a mathematically correct decision, it nullifies a potential advantage of mine. By opening to 2x, I can potentially a) win the blinds and antes, b) fold when a likely better hand shoves, or c) win the pot postflop with a c-bet or by flopping a strong hand. Either one of these 3 outcomes seem better than risking my tournament life based on marginal blocker value and equity stake. If I had a strong hand like AQ or even a hand with blocker/showdown value like A4s, I would probably just shove. With less than 15 big blinds, there is no room for even a 2x open and a shove is the only option, so KTs might just be a fold depending on the situation as you tighten your range and look for a spot to shove.
To put it simply, when you have less than 40-60 big blinds, you don’t want to risk 3+ big blinds as a preflop raise with hands that cannot call a 3-bet/shove. You also don’t want to only open hands that are strong enough to call a 3-bet/shove. Therefore a smaller open sizing is better.
Edit: You do need to increase your sizing if players have limped in front of you. 1x per limper may actually be too large when stacks are shallow (because you don’t want to be pot committed with 20 big blinds behind in a 14.5 big blind pot with AJ after you raise over limpers and get a caller). I typically use 0.5x per limper, but I am wary about opening over limpers because they like to call, so it is not as easy to open a weaker range of hands.