Prediction and Evaluation in College Hockey using the Bradley-Terry-Zermelo Model

We describe the application of the Bradley-Terry model to NCAA Division I Mens Ice Hockey. A Bayesian construction gives a joint posterior probability distribution for the log-strength parameters, given a set of game results and a choice of prior distribution. For several suitable choices of prior, it is straightforward to find the maximum a posteriori point (MAP) and a Hessian matrix, allowing a Gaussian approximation to be constructed. Posterior predictive probabilities can be estimated by 1) setting the log-strengths to their MAP values, 2) using the Gaussian approximation for analytical or Monte Carlo integration, or 3) applying importance sampling to re-weight the results of a Monte Carlo simulation. We define a method to evaluate any models which generate predicted probabilities for future outcomes, using the Bayes factor given the actual outcomes, and apply it to NCAA tournament results. Finally, we describe an on-line tool which currently estimates probabilities of future results using MAP evaluation and describe how it can be refined using the Gaussian approximation or importance sampling.

Prior Distributions for the Bradley-Terry Model of Paired Comparisons

The Bradley-Terry model assigns probabilities for the outcome of paired comparison experiments based on strength parameters associated with the objects being compared. We consider different proposed choices of prior parameter distributions for Bayesian inference of the strength parameters based on the paired comparison results. We evaluate them according to four desiderata motivated by the use of inferred Bradley-Terry parameters to rate teams on the basis of outcomes of a set of games: invariance under interchange of teams, invariance under interchange of winning and losing, normalizability and invariance under elimination of teams. We consider various proposals which fail to satisfy one or more of these desiderata, and illustrate two proposals which satisfy them. Both are one-parameter independent distributions for the logarithms of the team strengths: 1) Gaussian and 2) Type III generalized logistic.

Model-Based Cross-Correlation Search for Gravitational Waves from Scorpius X-1

We consider the cross-correlation search for periodic GWs and its potential application to the LMXB Sco X-1. This method coherently combines data from different detectors at the same time, as well as different times from the same or different detectors. By adjusting the maximum time offset between a pair of data segments to be coherently combined, one can tune the method to trade off sensitivity and computing costs. In particular, the detectable signal amplitude scales as the inverse fourth root of this coherence time. The improvement in amplitude sensitivity for a search with a coherence time of 1hr, compared with a directed stochastic background search with 0.25Hz wide bins is about a factor of 5.4. We show that a search of 1yr of data from Advanced LIGO and Advanced Virgo with a coherence time of 1hr would be able to detect GWs from Sco X-1 at the level predicted by torque balance over a range of signal frequencies from 30-300Hz; if the coherence time could be increased to 10hr, the range would be 20-500Hz. In addition, we consider several technical aspects of the cross-correlation method: We quantify the effects of spectral leakage and show that nearly rectangular windows still lead to the most sensitive search. We produce an explicit parameter-space metric for the cross-correlation search in general and as applied to a neutron star in a circular binary system. We consider the effects of using a signal template averaged over unknown amplitude parameters: the search is sensitive to a combination of the intrinsic signal amplitude and the inclination of the neutron star rotation axis, and the peak of the expected detection statistic is systematically offset from the true signal parameters. Finally, we describe the potential loss of SNR due to unmodelled effects such as signal phase acceleration within the Fourier transform timescale and gradual evolution of the spin frequency.

New Coordinates for the Amplitude Parameter Space of Continuous Gravitational Waves

The parameter space for continuous gravitational waves (GWs) can be divided into amplitude parameters (signal amplitude, inclination and polarization angles describing the orientation of the source, and an initial phase) and phase-evolution parameters. The division is useful in part because the Jaranowski-Krolak-Schutz (JKS) coordinates on the four-dimensional amplitude parameter space allow the GW signal to be written as a linear combination of four template waveforms with the JKS coordinates as coefficients. We define a new set of coordinates on the amplitude parameter space, with the same properties, which is more closely connected to the physical amplitude parameters. These naturally divide into two pairs of Cartesian-like coordinates on two-dimensional subspaces, one corresponding to left- and the other to right-circular polarization. We thus refer to these as CPF (circular polarization factored) coordinates. The corresponding two sets of polar coordinates (known as CPF-polar) can be related in a simple way to the physical parameters. We illustrate some simplifying applications for these various coordinate systems, such as: a calculation of Jacobians between various coordinate systems; an illustration of the signal coordinate singularities associated with left- and right-circular polarization, which correspond to the origins of the two two-dimensional subspaces; and an elucidation of the form of the log-likelihood ratio between hypotheses of Gaussian noise with and without a continuous GW signal. These are used to illustrate some of the prospects for approximate evaluation of a Bayesian detection statistic defined by marginalization over the physical parameter space. Additionally, in the presence of simplifying assumptions about the observing geometry, we are able to explicitly evaluate the integral for the Bayesian detection statistic, and compare it to the approximate results.

Treatment of Calibration Uncertainty in Multi-Baseline Cross-Correlation Searches for Gravitational Waves

Uncertainty in the calibration of gravitational-wave (GW) detector data leads to systematic errors which must be accounted for in setting limits on the strength of GW signals. When cross-correlation measurements are made using data from a pair of instruments, as in searches for a stochastic GW background, the calibration uncertainties of the individual instruments can be combined into an uncertainty associated with the pair. With the advent of multi-baseline GW observation (e.g., networks consisting of multiple detectors such as the LIGO observatories and Virgo), a more sophisticated treatment is called for. We describe how the correlations between calibration factors associated with different pairs can be taken into account by marginalizing over the uncertainty associated with each instrument.

Current status of gravitational-wave observations

The first generation of gravitational wave interferometric detectors has taken data at, or close to, their design sensitivity. This data has been searched for a broad range of gravitational wave signatures. An overview of gravitational wave search methods and results are presented. Searches for gravitational waves from unmodelled burst sources, compact binary coalescences, continuous wave sources and stochastic backgrounds are discussed.

Searching for Galactic White Dwarf Binaries in Mock LISA Data using an F-Statistic Template Bank

We describe an F-statistic search for continuous gravitational waves from galactic white-dwarf binaries in simulated LISA Data. Our search method employs a hierarchical template-grid based exploration of the parameter space. In the first stage, candidate sources are identified in searches using different simulated laser signal combinations (known as TDI variables). Since each source generates a primary maximum near its true Doppler parameters (intrinsic frequency and sky position) as well as numerous secondary maxima of the F-statistic in Doppler parameter space, a search for multiple sources needs to distinguish between true signals and secondary maxima associated with other, louder signals. Our method does this by applying a coincidence test to reject candidates which are not found at nearby parameter space positions in searches using each of the three TDI variables. For signals surviving the coincidence test, we perform a fully coherent search over a refined parameter grid to provide an accurate parameter estimation for the final candidates. Suitably tuned, the pipeline is able to extract 1989 true signals with only 5 false alarms. The use of the rigid adiabatic approximation allows recovery of signal parameters with errors comparable to statistical expectations, although there is still some systematic excess with respect to statistical errors expected from Gaussian noise. An experimental iterative pipeline with seven rounds of signal subtraction and re-analysis of the residuals allows us to increase the number of signals recovered to a total of 3419 with 29 false alarms.

Improved search for galactic white dwarf binaries in Mock LISA Data Challenge 1B using an F-statistic template bank

We report on our F-statistic search for white-dwarf binary signals in the Mock LISA Data Challenge 1B (MLDC1B). We focus in particular on the improvements in our search pipeline since MLDC1, namely refinements in the search pipeline and the use of a more accurate detector response (rigid adiabatic approximation). The search method employs a hierarchical template-grid based exploration of the parameter space, using a coincidence step to distinguish between primary (``true) and secondary maxima, followed by a final (multi-TDI) ``zoom stage to provide an accurate parameter estimation of the final candidates.

Prospects for Stochastic Background Searches Using Virgo and LSC Interferometers

We consider the question of cross-correlation measurements using Virgo and the LSC Interferometers (LIGO Livingston, LIGO Hanford, and GEO600) to search for a stochastic gravitational-wave background. We find that inclusion of Virgo into the network will substantially improve the sensitivity to correlations above 200 Hz if all detectors are operating at their design sensitivity. This is illustrated using a simulated isotropic stochastic background signal, generated with an astrophysically-motivated spectrum, injected into 24 hours of simulated noise for the LIGO and Virgo interferometers.

F-statistic search for white-dwarf binaries in the first Mock LISA Data Challenge

The F-statistic is an optimal detection statistic for continuous gravitational waves, i.e., long-duration (quasi-)monochromatic signals with slowly-varying intrinsic frequency. This method was originally developed in the context of ground-based detectors, but it is equally applicable to LISA where many signals fall into this class of signals. We report on the application of a LIGO/GEO F-statistic code to LISA data-analysis using the long-wavelength limit (LWL), and we present results of our search for white-dwarf binary signals in the first Mock LISA Data Challenge. Somewhat surprisingly, the LWL is found to be sufficient -- even at high frequencies -- for detection of signals and their accurate localization on the sky and in frequency, while a more accurate modelling of the TDI response only seems necessary to correctly estimate the four amplitude parameters.