Health Index - Possible Problem

This may or may not be intentional by design but I just don't know anymore so figured I should ask a question that I have had for a few years. I'm not asking Andy to disclose any formulas or inner workings of the game, I just would like to know if the math on the Health Index calculation is actually intentional or an error.

With the other main government indexes, their calculations follow a predictable and very simple math path, where Health bucks the trend. All I am asking is if it is intended, I am not calling for changes for the love of god please don't make some radical sudden change in response to this question or take offense to it being asked.

Here is a number by number example of what I am referring to:
Elementary Schools Needed - 35075
Elementary Schools on Hand - 76113
Elementary School Index - 217

Roads Needed - 290344
Roads on Hand - 586495
Roads Index - 202

Finance Income - 1.269T
Finance Cost - 1.071T
Finance Index - 118

Hospitals Needed - 2217
Hospitals on Hand - 5088
Health Index - 196

Everything besides the Health Index is an even on hand / needed, where the hospitals are actually at 2.29x needed but index is flat (very very slightly falling) at 196.

Thanks, auditor it was on my todo list, just not very high.

I suspect it's intentional, and is tied in some way to population age distribution. But, haven't gotten to checking that hypothesis.

It would have been sensible to bake in such a relationship so that if you got a lot old people, you need more healthcare. This something I would have expected to be there from the beginning.

If Andy doesn't respond, I will move it up my list of things to figure out.

It's obviously intended. This post is a repeat of the chat, with further calculation example.

After excluding, the outliers that have been losing or gaining hospitals at regular pace (eg, those where the health index is lagging due to moving average calculation)



As you can see (Have/Need) = (Health index) until 125 (regression says), Then above 125 it requires increasingly more hospitals to raise the index a defined amount.

I took a regression of the straight part of the line, and then transformed it to get the equation for calculating the required (have/need) to achieve a desired health index. Its complicated, and I'm not going to try to describe it unless someone realllly wants to know.

(have/need*100)=(desired Health index*1.3455)-43.38

Using Auditor's provided example, where 196 is the desired health index:

196*1.3455-43.38=220.3

It's off by ~9 for auditors example. I tried incorporating the given example into the regression, but it broke it. So either the example given isn't in steady state, or there may be some minor factor unaccounted for.

Further study, could say, but I think this answers the question.

-hymy-

After some further input (angry criticism). I went over to LU to diversify my dataset

Country #	Have	need	Have	Health	Delta
/Name			/Need	Index
1205	169	188	89.89361702	90.24	-0.346382979
3033	205	224	91.51785714	91.62	-0.102142857
2115	2807	2887	97.2289574	97.46	-0.231042605
1554	236	220	107.2727273	107.36	-0.087272727
Hymyland	711	580	122.5862069	122.65	-0.063793103
472	298	242	123.1404959	123.35	-0.209504132
LU701	1275	973	131.0380267	129.64	1.398026721
500	321	242	132.6446281	130.93	1.714628099
416	325	241	134.8547718	132.34	2.514771784
LU1019	1450	1062	136.5348399	133.8	2.734839925
460	350	242	144.6280992	139.78	4.848099174
N Hymyland	424	287	147.7351916	141.99	5.745191638
105	2035	1330	153.0075188	145.98	7.027518797
LU3464	6000	3752	159.9147122	150.72	9.194712154
LU1521	2599	1414	183.8048091	168.69	15.11480905
LU731	6283	3046	206.2705187	181.86	24.41051871
""	2860	1268	225.5520505	194.24	31.31205047
Auditor	5080	2226	228.2120395	196.2	32.01203953
LU1746	2500	990	252.5252525	210.9	41.62525253

After incorporating more data, it's possible that the relationship is actually a polynomial one. That Rsquare is better than the linear Rsquare.



It's easy to use that formula to make a range of Health indexes with calculated (have/need*100) numbers, plot it, and do another regression to get the needed equation. I don't have the right tools for this, so I'm in excel and I know there are rounding errors that could be avoided with something like MATLAB.



Sadly, the raw data comes from moving averages, because all the indexes in this game are that way. That results in uncertainty that could just look like a polynomial relationship. Use of the equations over time might allow some practical tweaking of the parameters.

The simple linear equation is (have/need/*100)=1.5812x-75.566 where x is the desired health index. The Rsquare is 0.9964

Either equation will not be exact, but will be close to the ball park. The linear one may work better for lower indexes, while the polynomial one may work better for the nosebleed section.

Forum malfunctioned.

The health index is not linear.
if you double the number of hospitals, the index will not double.

it is close to linear at low levels, it is getting further from linear as the numbers increase.

as far as I remember, this is true to more than one index at higher levels.