66 replies to this topic

[Snib]

pier4r, the problem is that XP on your crew is not displayed correctly, in particular your doubles not showing up. For example, looking at some of my fresh tanks, all crews started at 75% fresh (I never retrain):

Tank 1: 77% with 2510 XP, 79% with 2619 XP

Tank 2: 85% with 17508 to 19296 XP

Tank 3: 86% with 20513 to 20531 XP

Tank 4: 88% with 18303 to 18779 XP

Tank 5: 89% with 23838 to 23876 XP, 90% with 24590 XP

Should be immediately obvious that those numbers just do not work out, makes it kinda hard to evaluate the real curve.

[GoldenGnu]

I've made a small program that use this formula: WotXP

Also, I wrote down the XP I got for each battle on my SU-82, SU-100 and SU-152.
It match very well with the formula IMHO. If anyone is interested in the data, you can download it here (OpenOffice spreadsheet).
I did not write down when the crew was wounded. So, a bit of inaccuracy must be expected...

[pier4r]

Snib, on 08 September 2011 - 08:16 PM, said:

pier4r, the problem is that XP on your crew is not displayed correctly, in particular your doubles not showing up. For example, looking at some of my fresh tanks, all crews started at 75% fresh (I never retrain):

Tank 1: 77% with 2510 XP, 79% with 2619 XP

Tank 2: 85% with 17508 to 19296 XP

Tank 3: 86% with 20513 to 20531 XP

Tank 4: 88% with 18303 to 18779 XP

Tank 5: 89% with 23838 to 23876 XP, 90% with 24590 XP

Should be immediately obvious that those numbers just do not work out, makes it kinda hard to evaluate the real curve.

yep in fact considering an Xp of expShown*1.33 (or a bit more, like 1.5 to consider doubles) the formula is working perfectly. The chances are that this is the correct formula (or almost corrrect) are really high.
Now i'm @ 90% of crew skill (and the formula still working perfectly), started from 50%, other 10% and i post all the result.

Obviously it would be better if code snippet of experience calculation will be posted or, at least, correct stats are shown to the player.

[pier4r]

pier4r, on 10 September 2011 - 05:09 PM, said:

yep in fact considering an Xp of expShown*1.33 (or a bit more, like 1.5 to consider doubles) the formula is working perfectly. The chances are that this is the correct formula (or almost corrrect) are really high.
Now i'm @ 90% of crew skill (and the formula still working perfectly), started from 50%, other 10% and i post all the result.

Obviously it would be better if code snippet of experience calculation will be posted or, at least, correct stats are shown to the player.

Anyway in personal file of every crew member, i have tested it right now, the "accelerate experience" is shown, so my fix exp*1,33 is wrong. Now i should test the double experience for each victory.

edit: double exp for first victory seems to be not shown in personal crew stats.

[pier4r]

Ok, training complete!

column 1: skill
column 2: skill amount on personal file of the commander when i reached the skill
column 3: skill amount on column 2 multiplied for 1.35 (to compensate hidden stats, anyway "accelerate crew training" bonus on experience seems to be shown, while stats for the first victory of the day aren't shown, then the coefficent should not be 1.35 but less)
column 4: skill predicted by the formula

on a t26 that starts with elite status and accelerate crew training on.

50	0	0	0
51	371	501	524
52	691	933	1072
53	985	1330	1646
54	1843	2488	2247
55	2072	2797	2876
56	2428	3278	3536
57	2861	3862	4226
58	3400	4590	4949
59	4082	5511	5705
60	4510	6089	6498
61	5315	7175	7328
62	6077	8204	8196
63	6407	8649	9106
64	7236	9769	10059
65	7911	10680	11057
66	8896	12010	12101
67	9336	12604	13195
68	10241	13825	14341
69	10880	14688	15540
70	11919	16091	16796
71	12583	16987	18111
72	12923	17446	19488
73	13799	18629	20930
74	14685	19825	22440
75	15935	21512	24021
77	18193	24561	27411
78	20315	27425	29226
79	21133	28530	31127
80	22855	30854	33118
81	24297	32801	35202
82	25567	34515	37384
84	28539	38528	42063
85	30683	41422	44569
86	32983	44527	47193
87	35359	47735	49941
89	36897	49811	55831
90	38185	51550	58986
91	41493	56016	62289
92	44227	59706	65748
93	46483	62752	69370
94	47687	64377	73163
95	50005	67507	77135
96	52729	71184	81294
97	56477	76244	85649
98	58919	79541	90209
99	62587	84492	94984
100	65825	88864	99984

graph:


I think that the formula is quite correct, even if it, maybe, should be scaled to match the right shown values (to avoid calculations to compensate hidden stats)

Note that is obvious that initial boost to 75% of skill is quite useful and "cheaper".

here with real stats scaled by 1.51

50	0	0	0
51	371	560	524
52	691	1043	1072
53	985	1487	1646
54	1843	2783	2247
55	2072	3129	2876
56	2428	3666	3536
57	2861	4320	4226
58	3400	5134	4949
59	4082	6164	5705
60	4510	6810	6498
61	5315	8026	7328
62	6077	9176	8196
63	6407	9675	9106
64	7236	10926	10059
65	7911	11946	11057
66	8896	13433	12101
67	9336	14097	13195
68	10241	15464	14341
69	10880	16429	15540
70	11919	17998	16796
71	12583	19000	18111
72	12923	19514	19488
73	13799	20836	20930
74	14685	22174	22440
75	15935	24062	24021
77	18193	27471	27411
78	20315	30676	29226
79	21133	31911	31127
80	22855	34511	33118
81	24297	36688	35202
82	25567	38606	37384
84	28539	43094	42063
85	30683	46331	44569
86	32983	49804	47193
87	35359	53392	49941
89	36897	55714	55831
90	38185	57659	58986
91	41493	62654	62289
92	44227	66783	65748
93	46483	70189	69370
94	47687	72007	73163
95	50005	75508	77135
96	52729	79621	81294
97	56477	85280	85649
98	58919	88968	90209
99	62587	94506	94984
100	65825	99396	99984

[GoldenGnu]

I wrote down the xp I got after each battle (as I wrote in my previous post).
AFAIK that is most accurate way to collect the data...
IMHO the following proves the formula...

Here are some results:

Primary skill 75%-100%

Graph:


Data:

Spoiler                     

Level	SU-100	SU-152	Formula
75	0	0	0
76		1825	1581
77	3264	3607	3236
78	5343	5345	4969
79	6829	7704	6784
80	9587	8720	8684
81	10920	10952	10674
82	12895	13102	12758
83	15494	15825	14940
84	17776	17612	17225
85	20153	19849	19618
86	22568	22592	22123
87	26968	25155	24747
88	27958	27675	27494
89	30730	30616	30371
90	33493	33881	33383
91	36839	37789	36537
92	39989	40000	39840
93	43932	44621	43299
94	47509	47132	46921
95	51443	50996	50713
96	55501	55488	54684
97	59136	59684	58842
98	63689	63517	63196
99	68463	68868	67756
100	72722	74880	72530

First secondary skill 0%-100%

Graph:


Data:

Spoiler                     

Level	SU-85 (XP)	SU-85 (Rapid training compencation)	Formula
4	987	1234	427
6	1237	1546	672
11	1970	2463	1394
29	4694	5868	5930
30	5091	6364	6310
32	5838	7298	7124
33	6352	7940	7560
34	6692	8365	8017
35	7184	8980	8495
36	7454	9318	8996
38	8877	11096	10069
40	10083	12604	11246
41	10495	13119	11876
43	11857	14821	13227
44	12038	15048	13951
45	12964	16205	14709
48	14538	18173	17204
49	14956	18695	18116
50	16536	20670	19070
51	17235	21544	20070
52	18394	22993	21117
53	19036	23795	22213
54	19790	24738	23361
55	20687	25859	24563
57	22713	28391	27139
58	24250	30313	28519
59	25132	31415	29964
61	27257	34071	33061
62	28489	35611	34720
63	30539	38174	36457
64	31663	39579	38276
65	33947	42434	40181
66	34776	43470	42176
67	36397	45496	44265
68	38249	47811	46452
69	40182	50228	48742
70	42135	52669	51140
71	44022	55028	53651
72	47108	58885	56281
73	47625	59531	59035
74	51236	64045	61919
75	53256	66570	64938
76	55652	69565	68100
77	57721	72151	71411
78	61579	76974	74878
79	63977	79971	78508
80	67421	84276	82309
81	70828	88535	86290
82	74406	93008	90458
83	77154	96443	94823
84	80656	100820	99393
85	85042	106303	104179
86	88997	111246	109190
87	92579	115724	114438
88	97108	121385	119933
89	102640	128300	125687
90	106748	133435	131712
91	111783	139729	138021
92	116789	145986	144627
93	122743	153429	151545
94	128609	160761	158789
95	134368	167960	166374
96	140412	175515	174317
97	148117	185146	182634
98	155803	194754	191343
99	162093	202616	200463
100	169139	211424	210012

Notes
1) The Rapid training compensation is: XP + (XP/4) (Because the SU-85 got 4 crew members)
2) The spike at the end for SU-152 got it to 15% in the first secondary skill
3) The first secondary graph don't show all levels (because the first ~30 crew levels need so little xp - going up several levels a battle)
4) The raw data is available here (OpenOffice Spreadsheet)

[sword_of_Damocles]

nice calculation mate!!!!

[GoldenGnu]

@sword_of_Damocles
Thank you...
The raw data is still available here

EDIT:
All credits to barabaf for finding the formula - he/she must be really smart

[pier4r]

Golden gnu, on SU tanks you don't use rapid training, do you? In fact i need the compensation (like you, even if the better compensation will work, on a t26, with XP*1.5 instead of XP+XP/3 maybe because i play during operation avalance and i play during several days [double exp award]) like your camo and anyway how do your xp not need better compensation to follow the formula? So the "double exp for first victory" is correctly shown?

Anyway yes, i agree that the formula is not trivial and, for now, it's correct.

[GoldenGnu]

@pier4r
Because I wrote down the xp I got from the battle report after each battle and did not look at the stats...
Stats are wrong - crazy monkey wrong!

Look at raw data (linked above), and you'll see what I did...

I only used rapid training on my SU-85 (because it owns).
I did not use rapid training on the SU-100 and SU-152. (I moved to the next tank when everything was researched)

Operation avalance or simply x2 XP first victory of the day, will make your stats more and more inaccurate...

All stats regarding XP are wrong as soon as you have had a x2 XP victory: Total, Average, and Maximum (in crew/tank/total stats)

I hope my reply makes it clear why the formula work and the in-game stats does not...


EDIT:
Maximum XP, might still be correct after a X2 XP victory (theoretical)

[pier4r]

GoldenGnu, on 16 September 2011 - 01:37 PM, said:

@pier4r
Because I wrote down the xp I got from the battle report after each battle and did not looking at the stats...
Stats a wrong - crazy monkey wrong!

Ah! now i got it, yep i agree with your method.

[GoldenGnu]

@pier4r
Good!  

[Hagbardur]

I'd +1 everyone in this thread for nice detective work if I had any + left to give!

Love this kind of topics.

[GoldenGnu]

@Hagbardur
IMHO just +1 the original post by barabaf - he/she deserve all the credit!

[GoldenGnu]

I just wanted to post this, because it's not quite the numbers barabaf posted...
I think this is more accurate. It's just my humble opinion, though.

Primary 50-100
95.458 XP

Primary 75-100
72.530 XP

1st Secondary 0-100
210.012 XP

2nd Secondary 0-100
420.076 XP

3rd Secondary 0-100
840.209 XP

From Primary 50 to 3rd Secondary 100
1.565.755 XP

See more in my program: WotXP  

[Walzwerk]

Thanks for your impressive work!
There is one things, that puzzles me, though:
I am training the crew of my SU-85 right now.
Is it true that different crewmen have different limits for their primary qualification percentages?
I´m only asking, because the driver was the first one to hit 100%, although he had the lowest xp count of all four (46.637xp, compared to 46.934 of the loader, who is still at 99%).
How is that possible?

Thanks for any answer, Cheers!

[attachment=16561:CrewTrainingBW.jpg]

[Snib]

Walzwerk, on 27 September 2011 - 12:40 AM, said:

How is that possible?

As stated a couple times in this thread, the crew XP count shown in game is unfortunately incorrect.

[ThulePhil]

All your calculations might be correct for the basic Skill but not for the whole.

I have a T-54 with almost 1300 battles.

This tank has reached 100-100-100-100 long time ago.

I have additional 400.000 XP on it. The total XP of the tank is 1.276.000. The crew is at 1.700.000. Do the math

[GoldenGnu]

@Walzwerk
When the crew members are wounded they'll get a XP penalty.
So, the crew member that was wounded the least will be the first to hit 100%.
Also as Snib said: Stats are wrong...

@Snib
QFT

@ThulePhil
I assume you have read the entire thread? remember that the in-game stats are wrong?
Please: Share your valid data (AKA not your in-game stats), that proves the formula wrong

[Walzwerk]

GoldenGnu, on 27 September 2011 - 08:23 AM, said:

@Walzwerk
When the crew members are wounded they'll get a XP penalty.
So, the crew member that was wounded the least will be the first to hit 100%.
Also as Snib said: Stats are wrong...

Snib, on 27 September 2011 - 01:00 AM, said:

As stated a couple times in this thread, the crew XP count shown in game is unfortunately incorrect.

Thank you very much for your answers, now I think I got it. I didn´t think about the penalty. I thought, the deviation from the real numbers would affect all crewmembers equally, and penalty would be reflected in lower xp count of the crewman, which seems not to be the case. Obviously, I didn´t take down notes with pen and paper....
Thanks again, Cheers!

P.S.: Unfortunately, I don´t have positive votes left. I hope, I´ll remember tomorrow...