Frankly, the predictability depends on the length of the format.
Tests are least predictable because the length of the format allows time to recover from the most disastrous of situations.
ODI's are more predictable only because a disastrous response of the team batting second or a great reply after a disastrous performance by the team batting first, makes a recovery virtually impossible. The same should theoretically apply to still shorter versions. 220 runs by the team batting first and 60 for 7 in reply in the first ten overs, kills the game, and thats a fact as does 110 by the team batting first and 80 for 2 in 10 overs in response.
We have just not seen enough of T20 matches yet for all possible scenarios to play out and repetitions of matches that have gone before start hitting us in the face.
That is why the non-cricketing aspects viz cheer-leaders, orange caps, awards for most sixes, length of hits, number of sixes. and general emphasis on the boundary and over the boundary hit being what constitutes entertainment rather than the cricket per se is stressed so much.
Also the hype of city loyalties and fans versus fans (most IPL ads are based on this) is expected to keep the fans interested by sheer strength of their loyalties and the emotions attached.
The jury, I believe is still out on whether this will, in the long term sustain interest even if the cricket is not great, is still out.
Otherwise, if you just realise that a run a ball chase of the odi's is the same as a 9 runs per over in the T-20 format (with 180 runs in T20 being like 300 in odi's) you can extrapolate most scenarios and what constitutes economical bowling or a good strike rate.
Batsmen and bowlers have still not settled down with the 'strategies' of how to 'structure' an innings which is what led to the so called middle-overs-boredom of the 50 over format. Once this has been worked out over the next couple of years, T20 matches will also follow a "repetitive-model."
This is built into the 'shortness' of the game where maximising run rate (which may lead to risk taking' has to be balanced against early loss of wickets which leads to prudence.
The only way that can be avoided is where one of them, say loss of wickets, becomes irrelevant. To put it in a ridiculously exaggerated way (just to put the point across), it would not matter if ten balls or less constituted an innings. Then every ball could be hit (tried to be hit) for a six.
The problems of odi's middle overs are partly self created by the field restrictions of the initial overs which meant that hitting was reserved for the beginning and the death. They tried to overcome this by the flexibility of the power play but most captains refused to make much use of this flexibility and tried to get this 'risky' period out of the way and done with at the earliest so as to be able to focus on the more traditional 'strategies' while leading a fielding side.
As in the economy, so in every other sphere, too much interference without rectifying the basic flaws never leads to long term solutions that work.
Test cricket (and first class cricket) was having problems of spectator indifference primarily due to the boring draws of the fifties and sixties. Instead of modifying these longer games to ensure more results at that time, MCC came up with the one day tournaments at the county level and after the initial success these became international too. But it hasn't worked in the long run because the original issue was NOT tackled.
What it did, however, was to introduce a form of the game which took much less time BESIDES bringing assured results, and it has become that much more difficult to bring the crowds back to the longer version now because the spectators are just not used to watching cricket for 'days on end'