COVID-19 short-term forecasts Confirmed 2020-03-24


Disclaimer

  • Forecasts produced by Jennie Castle, Jurgen Doornik, and David Hendry, researchers at the University of Oxford. These are our forecasts, and should not be considered official forecasts from, or endorsed by, any of: University of Oxford, Oxford Martin School, Nuffield College, Magdalen College, or any other University of Oxford institute.
  • These forecasts are short term time-series extrapolations of the data. They are not based on epidemiological modelling or simulations. The documentation that is provided is still in progress and not peer reviewed. All forecasts are uncertain: their success can only be determined afterwards. Many mitigation strategies are in place, which, if successful, invalidate our forecasts. An explanation of our methods is provided below.

Moderation of forecast

[2020-03-24] Our forecasts are starting to overestimate in some cases. This was always expected to happen when the increase starts to slow down. We have started to add forecasts that are based on what happened in China earlier this year, but only for Italy and Spain sofar.
[2020-03-26] We are in the process of updating our forecasts to reflect China's experience in a more coherent manner. This is shown for Italy only, and will require further development.
[2020-03-29] Now including some US States, based on New York Times data.

Further information

  • We believe these forecasts fill a useful gap in the short run. They give an indication of what is likely to happen in the next few days, removing some aspect of surprise. Moreover, a noticeable drop in comparison to the extrapolations could be an indication that the implemented policies are having some impact. It is difficult to understand exponential growth. We hope that these forecasts may help to convince viewers to adhere to the policies implemented by their respective governments, and keep all arguments factual and measured.
  • We use the data repository for the 2019 Novel Coronavirus Visual Dashboard operated by the Johns Hopkins University Center for Systems Science and Engineering. This is updated daily, but it seems that our forecasts need slightly less frequent updating.
    US state data as of 2020-03-28 is courtesy of the New York Times. This seems to be a day behind the Johns Hopkins data.
  • We can only provide forecasts of what is measured. If confirmed cases are an underestimate of actual cases, then our forecasts will also be underestimates. Again: no other epidemiological data is used.
  • We will probably revise or update the methodology as we learn what is happening in the next few days or weeks.
  • Once the number of cases levels off, there is no need to provide these forecasts anymore.
  • Countries where the counts are very low or stable have been omitted.
  • We will probably revise or update the methodology as we learn what is happening in the next few days or weeks.
  • Once the number of cases levels off, there is no need to provide these forecasts anymore.
  • Countries where the counts are very low or stable have been omitted.
  • The graphs have dates on the horizontal axis (yyyy-mm-dd) and cumulative counts on the vertical axis. They show
    1. bold dark grey line (with circles): observed counts (Johns Hopkins CSSE);
    2. many light grey lines (with open circles): forecasts using different model settings and starting up to four periods back;
    3. red line (with open circles): single forecasts path using default model settings;
    4. black line (with crosses): average of all forecasts, recentered on the last observation;
    5. thin green lines: some indication of uncertainty around the red forecasts, but we do not know how reliable that is.
    Both the red line forecasts and the black lines are also given in the tables above. These forecasts differ, we are currently inclined to use the average forecasts.
  • The forecasts are constructed as follows:
    1. An overall `trend' is extracted by taking a window of the data at a time. In each window we draw `straight lines' which are selected using an automatic econometric procedure (`machine learning'). All straight lines are collected and averaged, giving the trend.
    2. Forecasts are made using the estimated trend, but we note that this must be done carefully, because simply extrapolating the flexible insample trend would lead to wildly fluctuating forecast. We use the `Cardt' method, which has been found to work well in other settings.
    3. Residuals from the trend are also forecasted, and combined with trend forecasts into an overall forecast.
  • EU-BS is Estonia, Latvia, and Lithuania together.
  • This paper describes the methodology and gives further references, but is still preliminary.

Confirmed count forecast average (bold black line in graphs) 2020-03-25 to 2020-03-29

DateUKEUATBEBSCZDEDKESFIFRGRHRIEITNLPLPTROSESINOCH
2020-03-24 8077 195234 5283 4269 1677 1394 32986 1591 39885 792 22304 743 382 1329 69176 5560 901 2362 794 2286 480 2863 9877
2020-03-25 9400 221000 6120 4780 1900 1540 37300 1690 45800 880 25400 810 450 1530 76000 6270 1030 2720 930 2480 520 3120 11100
2020-03-26 10700 252000 6880 5330 2110 1680 42000 1800 52500 960 28800 860 510 1730 84000 7000 1140 3100 1010 2670 550 3370 12600
2020-03-27 12300 286000 7750 5960 2350 1840 47400 1920 60300 1050 32600 920 580 1950 93000 7800 1260 3550 1110 2870 580 3640 14300
2020-03-28 14200 326000 8740 6660 2620 2010 53400 2040 69200 1150 37100 980 660 2210 104000 8700 1400 4060 1230 3090 620 3930 16300
2020-03-29 16300 370000 9880 7440 2920 2190 60200 2170 79600 1250 42100 1050 760 2500 115000 9720 1560 4650 1350 3330 660 4250 18500

Confirmed count forecast average (bold black line in graphs) 2020-03-25 to 2020-03-29

DateAustraliaBrazilCanadaIranMalaysiaPhilippinesSouth AfricaUS
2020-03-24 2044 2247 2790 24811 1624 552 554 53740
2020-03-25 2370 2690 3300 26500 1780 630 660 66000
2020-03-26 2710 3180 3760 28300 1930 700 730 80000
2020-03-27 3100 3760 4290 30200 2090 780 810 97000
2020-03-28 3540 4440 4910 32300 2270 860 900 117000
2020-03-29 4050 5260 5640 34500 2460 960 1010 143000

Confirmed count forecast (bold red line in graphs) 2020-03-25 to 2020-03-29

DateUKEUATBEBSCZDEDKESFIFRGRHRIEITNLPLPTROSESINOCH
2020-03-24 8077 195234 5283 4269 1677 1394 32986 1591 39885 792 22304 743 382 1329 69176 5560 901 2362 794 2286 480 2863 9877
2020-03-25 9300 219000 6000 4630 1890 1510 36900 1680 45600 900 24600 790 430 1480 76000 6020 1010 2530 850 2490 520 3110 10800
2020-03-26 10800 244000 6900 5090 2130 1640 41700 1790 52200 1010 27500 840 490 1650 83000 6670 1150 2780 970 2720 560 3400 11900
2020-03-27 12500 272000 7900 5600 2400 1780 47000 1900 59900 1140 30800 900 560 1850 90000 7380 1310 3060 1110 2980 600 3710 13100
2020-03-28 14400 304000 9100 6150 2690 1930 53100 2030 68600 1290 34400 970 630 2060 99000 8170 1480 3350 1270 3250 640 4040 14500
2020-03-29 16700 338000 10400 6760 3020 2090 60000 2170 78700 1450 38500 1040 720 2290 108000 9040 1680 3680 1460 3560 690 4410 16000

Confirmed count forecast (bold red line in graphs) 2020-03-25 to 2020-03-29

DateAustraliaBrazilCanadaIranMalaysiaPhilippinesSouth AfricaUS
2020-03-24 2044 2247 2790 24811 1624 552 554 53740
2020-03-25 2340 2600 3180 26500 1740 640 620 65000
2020-03-26 2690 3030 3770 28500 1880 730 720 79000
2020-03-27 3090 3520 4490 30600 2030 830 850 96000
2020-03-28 3540 4080 5320 33000 2200 950 990 116000
2020-03-29 4050 4730 6310 35500 2380 1090 1150 141000

Initial visual evaluation of forecasts of Confirmed