You cannot teach an old dog new tricks. Or so they say. Well, just try telling Winston Peters that.
The master conjurer not only pulled a couple of rabbits out of his top hat this week. He had skinned and gutted them before you could cry "Hone Harawira" or "Te Ururoa Flavell".
In ruling out working with Internet-Mana because it is a "race-based party", Peters has effectively pulled the plug on his party's participation in what at best would be an unwieldy four-party government and at worst a highly unstable one.
Peters has judged that being the kingmaker responsible for what would be a left-dominated government would be the death of NZ First.
David Cunliffe might have kept Peters on board by ruling out any working arrangement with Harawira and friends. But he refused to do so, perhaps in the vain hope that Labour, the Greens and Internet-Mana might win just enough seats to cut NZ First out of the power equation.
That is dreaming. The reality is more the stuff of nightmares - most notably Labour's crumbling poll ratings which are responsible for the mess on the left and into which NZ First finds itself being dragged.
Peters' ban on his party working with Internet-Mana has also quietly dealt with another dilemma.
While Peters could have formed a government with Labour alone were the latter's support to lift significantly at next month's general election, such an arrangement would still have most likely had to have been propped up by some Greens feeling grumpy at being shut out of the action.
Again, instability beckoned. Not so on the right, however. Peters has applied the same working ban to the Maori Party. Because the latter is likely to be wiped out as a parliamentary force on election day anyway, the impact on National will be negligible.
And there lies the rub. Peters' stance on working with "race-based" parties has wide-ranging implications which are not immediately obvious.
Similar significance applies to this week's other major political development: John Key's decision not to gift an electorate seat to Colin Craig and his Conservative Party.
The separate announcements might seem unconnected. But if you join the dots, you reach an unavoidable conclusion: unless National can achieve what is considered to be the near impossible and muster enough votes to be able to govern alone, then the next Government is odds on to be a National-NZ First one. What is more, both Peters and Key seem to accept that the logic of such an arrangement is becoming inescapable. The details can be sorted out much later.
Key's demeanour towards Peters has certainly undergone a change. Key assumed Peters would punish him for refusing to include Peters in post-election negotiations in 2008 and 2011. The Prime Minister instead seemed to take special delight in poking a stick through the bars of Peters' cage and ruffling the 1950s New Zealand that was hidden inside and which NZ First seeks to recreate at large.
Such taunts are of no help in normalising relations between NZ First and National.
The Prime Minister has noticeably backed off this practice as the election has neared.
More than that, Key has now ensured Peters is the clear winner from National's decision not to relinquish Murray McCully's East Coast Bays electorate, thereby giving the NZ First leader a huge advantage when it comes to soaking up the votes of provincial and back-block conservatives who have yet to be convinced the economy is on the up or who bristle at Key's attempts to appeal to metropolitan-based liberals.
Peters will secure these votes because those who hold them know NZ First will clear the 5 per cent threshold, whereas Craig's party probably will not.
Key has not done Peters this favour out of any feelings of kindness.
He has done it because it obliges Peters to negotiate with him first post-election because that conservative chunk of the vote expects Peters to serve as a brake on National.
The last thing those voters want to see is Peters installing a Labour-Greens-Internet Mana administration.
Key has also done it in the hope that NZ First can find a permanent niche and thus provide National with a long-term ally able to deliver at least a handful of seats to keep the centre-right in power.
While erecting his standard brick wall to specific questions regarding this scenario, Peters has dropped sufficient hints to suggest he is not averse to Key's strategy.
Peters has little choice in the matter, however.
Labour's crumbling poll ratings have cost Peters the leverage over Key and Cunliffe he previously seemed well positioned to enjoy.
Peters' longstanding gambit of refusing to divulge before the election which of the two major parties he would prefer as a partner in a coalition or some other government support arrangement has been rendered worthless.
Peters knows that Key knows Peters can no longer play him off against Cunliffe because the latter is staring down the barrel of a defeat that may be even worse than the horrendous result in 2011.
The rolling average of the various public polls now has support for Labour slumping to just under 26 per cent.
Victory would require both NZ First and Internet Mana to clear the 5 per cent threshold and pick up one or two percentage points on top of that, but not by siphoning off any more votes from Labour or the Greens.
Furthermore, all this requires that support for National drops below 47 per cent - the level the governing party recorded last time - as well as the Conservatives failing to clear the threshold.
Things get a little easier for the centre-left if Labour can claw its way back to the 30 per cent mark. But it would still require good showings by both NZ First and Internet-Mana.
An unlikely Labour win requires that every variable in these complex equations comes up trumps.
Fairy tales and political reality are rare bedfellows, however.
Even then Peters would have to give National - as the largest polling party - first shot at negotiating a deal.
And events are conspiring to make it ever more likely that Key and Peters will strike one if need be.