It isn't as easy as move all double indices. Each index must be considered separately considering how much advantage is at the index, how quickly advantage accumulates after the index is exceeded, and bet size at or near the index. It is a stricter restriction than RA indices but there needs to be a method to making the decisions. Personally at large bets I want a bigger gain in EV than with little bets. It really comes down to personal preference. Do some research into it. Don't trust what anybody you don't know says. Check it out. You give up some win % to double unless you would only take 1 card anyway. Selling the ability to take another card which will result in a lower W/L% needs to be viewed a bit different than when you would take just one card no matter what. In the latter situation you win and lose hands with 100% correlation whether you double or not. So doubling becomes strictly a question of volatility if you know you will only take one card. Volatility depends on bet size. How much extra EV is worth the added volatility? For any hand verse a particular dealer up card the answer may be no the added EV is not enough to make me want the extra volatility but at some point the extra EV from doubling is too much to ignore. How much EV depends on the size of your bet and the percentage gain.
How do they manage that when it is one of the players who cuts the shoe?
Good explanation. I was talking about selectively using more aggressive risk aversion than the risk averse index criterion set by Don and usually referred to as RA indexes. Frequencies at the index are important and where you are on the bellcurve is important. If the index is negative, moving the index higher has a small impact on the hand verse dealer card overall doubling frequency (how much you alter the percentage of the time the hand verse dealer up card is doubled), the bet will likely be small and the combination of the two will have little impact on volatility. It is probably doesn't affect volatility enough to change the index. At a small nonnegative index there is a lot of frequency at the index relative to the frequency further above the index and bets may be a bit bigger. This could be worth an index adjustment if the EV gain at the index is small enough. But at larger positive indices the change in the percentage of time you double by moving the index becomes more relative to the rest of the time you double because of the shape of the bell curve. You are in the most vertical part of the bell curve and eliminating the most frequent TC sector while bets are rapidly getting bigger as the TC increases. This combination makes it the most impactful on volatility to be more patient. Your highest indices are very infrequent and in the tail of the TC frequency bell curve. but have the highest bets out when they are used. You should certainly consider RA indexes there but their impact on volatility will be less dramatic if all things are equal. So you basically there is a limited number of doubles to consider and you know where to take the closest look at being patient to double and assess the cost of doing so. The rate of gain is a good indication of your costs correlation to the playing EoRs for any hand verse a dealer hand. If the correlation is poor you can either do something to make it stronger if you track enough information, or move the index.
I'm asking specifically about how they bring the small cards to the front of the shoe when they don't know where it will be cut?
No, the shoe would have to be cut at the start of the small clump. The small clump could just as easily be cut to the middle, the end, or out of play entirely.
This is a choice that is made by APs. You must decide if you really believe crazy swings with less certain BR growth is worth getting a little extra EV when your EV is already huge. I have always thought big BR players think about a lot of things backwards because they think their BR will allow them to be close to EV over time without risking busting out. It is true on the latter but not the former. The law of large numbers just tells you the difference between expectation and actual results will become more insignificant when compared to the total, not that it will get closer to expectation in terms of dollars. The swings associated with EV maximizing are brutal. You can do something about that without spending much EV if you do it surgically. For players that can raise their bets it may even be an EV gain when you bet to the same RoR but most with huge BRs players can't or don't want to bet larger bets for heat or table limit reasons. This causes them to not see any value to the change because it losses EV. But the gain is still there in reduced variance (volatility) and is still just as valuable. Remember the strength of you play is about both EV and variance as Don is so famous for his method of quantifying the trade off with SCORE. Everyone should understand that. Variance is the likelihood that successive rounds will be resolved the same. When playing two spots, covariance tells the likelihood that you will get a win/loss push on a round. Lower of each is better. But why is it better?