The trick is that a weak decay remembers which way its parent was pointing. When a J/psi produces a Lambda and an anti-Lambda, the two emerge with entangled spins; each then decays, the Lambda to a proton and a pion, the anti-Lambda to an antiproton and a pion, and the angles of those daughters carry information about the parent's spin direction [11]. Each side is therefore its own polarimeter. A simultaneous fit to all of those angles across about three million clean events makes the spin correlation measurable, and an electric dipole moment shows up as a small distortion of that pattern [2][11].
That design exists because the usual apparatus is unavailable. The tight EDM limits on electrons, neutrons, atoms and molecules come from holding a spin in an external field and watching it precess, which a particle that decays almost immediately cannot do, and that is why the Lambda's only previous direct number came from a fixed-target experiment at Fermilab more than four decades ago [6][5].
Three orders of magnitude is quick to write and worth unpacking. Running it backwards from the 10^-19 e cm now reached puts that four-decade-old direct limit near 10^-16 e cm, and the gap between the two is the parameter space this null result closes [4][13]. Statistics alone would not have covered it: uncertainty in an angular fit falls as the square root of the sample size, so a thousand-fold sensitivity gain from counting would take on the order of a million times the events [14]. The phys.org report does not give the Fermilab measurement's event count [12], so the split between better technique and bigger sample cannot be read off it.
What 10^-19 e cm does not tell you is how far the search sits from a decisive test, since no Standard Model expectation for this quantity appears in the account, which also puts no numbers beside the electron, neutron, atom and molecule constraints it calls complementary [12]. The case for hyperons is about coverage. A permanent EDM would mean a particle's charge distribution has a preferred orientation relative to its spin, violating time-reversal symmetry and, if CPT holds, CP along with it [7]; the Standard Model's known CP-violating sources are too small to explain why the observable universe kept its matter [8]. If the missing source is flavor-dependent, its pull on a strange quark need not match its pull on up and down quarks, which is why BESIII presents a Lambda constraint as complementary to the neutron, atom and molecule searches rather than as beating them [9]. For anyone building models that hang the new phase on the strange quark, the usable output here is a number to fit under, not a signal to explain [3][4].
