# Furness Method

Some competitors may profess to use "mathematical models" (such as 'Furness' etc.) to "count" a roundabout - whilst others will not declare this fact, but use the method in any case. It will be shown further below that no such "count" is actually possible using a mathematical model.

TSUK does not employ any form of modelling to derive flows. TSUK provides a 100% true and accurate count, either from the observed turning counts (where those counts can be physically seen) or using number plates (where turning counts cannot be seen because of trees in the middle of a roundabout, for example). Whilst mathematical models serve a useful purpose as far as estimates are concerned, most traffic surveys are not normally estimating exercises - clients purposefully and rightfully expect a full and true count.

A popular method is the 'Furness' method, as applied to a 4-arm roundabout. The main premise of the 'Furness' model is to use a very limited number of known movements and volumes, and then to algebraically estimate numeric quantities for the remaining unknown movements and volumes. As such, for each arm of a 4-arm roundabout, the model contains only 3 true values (only one of which is actually a turning movement): the first respective left from each arm; the total of all the "Ons" from each arm to the roundabout; and, the total of all the "Offs" to each arm of the roundabout.

Hence, for each arm, there remain 3 unknown parameters: 2nd left (or "2nd exit" or otherwise "ahead"), 3rd exit (or frequently "right"), and the inevitable U-turn (irrespective of significance). This means a total of 12 unknown parameters.

Therefore, using the 'Furness' method for a 4-arm roundabout, where there can be 16 possible movements, only 4 are actually known (i.e. the first left-turn from each arm).

Using all of those known and unknown parameters, a series of algebraic equations are setup, and the ensuing procedure is then to (forcibly) "balance" those equations to yield values for the unknown parameters. In this procedure, nothing is said about the ACTUAL O-D pairs.

Furthermore, in such models, U-turners are assumed to not exist. In that case, those who produce/use such models, deliberately insert random numbers in U-turning movements - which then requires the producer of the report to deliberately adjust numbers on other movements to ensure that the whole report of the counts (or system) balances. At face-value, appearance of U-turners within the report has the effect of bolstering the apparent authenticity of the "turning count" report.

So, the modelled result does genuinely appear to balance - i.e. for any given time segment, such as 15 minutes, all the traffic entering the roundabout is equal to all the traffic leaving the roundabout. This is only made possible because the total "Ons" and "Offs" are known and have been used as the fixed constants. However, the derived O-D pairs (apart from the 'First Left') are not true and will never be the same as a PROPER AND FULL turning count (either as actually observed or using number plates).

In summary, clients are normally not aware of the covert use of this method. Some survey companies deliberately use this procedure because it is quick and cheap (as far as job costs are concerned). In effect, where the correct approach is to count all 16 turning movements for a 4-arm roundabout, only 3 are actually counted. Hence, it only costs 1/5 of the would-be cost, and the result is then supplied to the client as a "turning" count, although it is patently not.

TSUK only provides 100% true and proper turning counts - backed-up by video evidence.

### Alternative Counting Method

We wish to draw the attention of the Client to the fact that some companies deliberately use covert methods which produce results that appear to be “true” and/or “representative”. This is especially so when such “alternative” methods are used to analyse car parks and roundabouts.

For car parks, some firms choose only to take the number plate of specific vehicles (such as a specific colour), and then use that result as a factor for the total volume of cars entering and leaving a car park.

For roundabouts, a popular method is the application of the ‘Furness’ model. This makes use of a limited number...