David Barnett's Brand Protection Articles: Unregistered Gems Part 5: Using groupings to find brandable domains

Introduction

The initial instalment of my recent series of articles on domain name discovery^[1] considered the use of phonotactic analysis - that is, the measure of the similarity of a string to the 'corpus' of other words in a language - to identify available unregistered candidate domains which may be of interest for potential brandability. This filtering process is necessary because of the large universe of domains which must be assessed. Considering just 5-character alphabetics .com domains, for example, there are approximately 9 million unregistered combinations of characters (as the SLD, or second-level domain name - i.e. the part to the left of the dot), out of the 'pool' of around 12 million possible names (from aaaaa.com to zzzzz.com).

Phonotactic analysis is a powerful tool, but does have some shortcomings - not least, it is computationally slow to calculate the phonotactic 'violation score' for a string of characters, but additionally it typically still retains large numbers of candidate domains within any given score window, and furthermore the 'mapping' of score to brandable desirability is not always 'clean' (i.e. many domains which are (subjectively) attractive do not always generate low violation scores).

In this follow-up, I start to explore additional frameworks for filtering the large sets of candidate domain names, considering the inherent structure of the SLD strings themselves. This type of approach would allow would-be brand owners to specify a preference as to the 'type' of brand name they may be looking to use, based on analogy with other brand names, words or strings (and potentially also allows for further filtering based on factors such as preferred initial letters, etc.), but without having to specify a specific exact string or keyword which they would like the brand name to resemble (as in the methodology proposed for 'variant string' domain names in another recent study^[2]).

Framework

As was also the case for the phonotactic method, the framework considered in this initial study relates to classification of domain names according to their high-level phonetic characteristics, but using a much simplified approach (and with negligible computational overhead to calculate) in which the constituent characters (consonants / vowels) are categorised into groups.

The groupings are based on the standard classifications for (English) consonant phonemes (i.e. unique sounds), in which they are classified according to the positions (in the vocal tract) and manners of articulation, within speech^[3,4].

For simplicity, I consider one of the original datasets utilised in the initial study - that is, the set of (as of the time of original analysis) unregistered, 5-character .com domain names with SLDs of the form CVCVC (C = consonant, V = vowel, noting that a 'y' is also accepted where it appears in a 'vowel' position). In general, there is no one-to-one mapping between individual characters and phonemes, due to factors such as variabilities in pronunciation and the existence of character combinations (especially phonemes such as 'ch', 'sh', 'ng', etc.). However, the use of the CVCVC pattern means that a 'cleaner' mapping can be assigned (since, for example, no consonant pairs will arise) and the utilised groupings are shown in Table 1. The overall classification of any given SLD string is then based just on the consonants present within the string (which are deemed to determine the high-level 'structure' of the word).

Top-level group	Group	Type	Consonants
1 (plosive)	1A	Bilabial plosive	b, p
1 (plosive)	1B	Alveolar plosive	d, t
1 (plosive)	1C	Velar plosive	c, g, k, q
2 (nasal)	2A	Bilabial nasal	m
2 (nasal)	2B	Alveolar nasal	n
3 (fricative)	3A	Labiodental fricative	f, v
3 (fricative)	3B	Alveolar fricative	s, x, z
3 (fricative)	3C	Glottal fricative	h
4 (approximant)	4A	Labial-velar approximant	w
4 (approximant)	4B	Retroflex approximant	r
4 (approximant)	4C	Palatal approximant	y
5 (lateral approximant)	5A	Alveolar lateral approximant	l
6 (affricate)	6A	Postalveolar affricate	j

Table 1: Groupings assigned to individual consonants as used in the analysis

In order to give a less granular framework, and cluster together words comprising (at a higher level) similar 'types' of sounds, the phoneme groups are combined into 'top-level groups' based on their manners of articulation (plosive, nasal, etc.), as shown in Table 1.

This methodology means that all CVCVC strings can be defined according to the top-level groups of their three consonants (i.e. characters 1, 3 and 5) and thereby expressed as a three-digit 'code' (the 'word type'). For example, the brand 'rolex' would be assigned word type 435 (since 'r' appears in consonant group 4, 'l' in group 5, and 'x' in group 3). In total, therefore, there are 216 (i.e. 6 × 6 × 6) distinct possible word types (from 111 to 666).

Analysis

Utilising the above framework, the 137,648 unregistered CVCVC .com domain names from the original study can be categorised into these 216 groups, based on their SLDs. As an initial round of analysis, it is informative to assess whether the domains assigned to each of the specific 216 word types tend (on average) to be associated with lower or higher phonotactic violation scores (as per the original study) - i.e. whether they comprise more or less 'credible' candidate words / brand names. This analysis is shown in Appendix A.

The data can also be visualised by plotting the spread of phonotactic violation scores of the domains within each group (Figure 1, in which the word types are sorted into their ranking order (Appendix A) from lowest ('best') to highest ('worst') mean phonotactic violation score - e.g. rank 1 = word type 225, etc).

Figure 1: Spread of phonotactic violation scores of the domains within each word-type group (sorted from lowest to highest mean phonotactic violation score (shown in red))

Examples of the individual domain SLDs (i.e. candidate brandable names) within each of the top four (i.e. lowest mean phonotactic violation score) groups are listed below for illustration (with their phonotactic scores shown in brackets in each case) (full lists shown in Appendix B).

Word type 225 - memyl (0.00), munyl (1.63), nemul (0.00), nenyl (0.00)
Word type 525 - lemyl (0.00), limyl (1.05), lunyl (0.93), lynul (0.00)
Word type 555 - lalyl (0.90), lelyl (0.00), lylel (1.05), lylul (1.05)
Word type 125 - banyl (0.12), cemyl (0.00), dimyl (0.00), gamyl (0.70)

Whilst in these specific cases, the actual strings may not, in general, be enormously compelling as brandable examples, the approach does broadly seem to behave in the intended way (i.e. in allowing the extraction of groups of 'similar' names to be reviewed in bulk, without relying on phonotactic score, whose behaviour is perhaps less predictable); the key 'art' to this approach might be in determining which word types tend to make the 'best' brandable candidates (and this may not necessarily always be those with the lowest phonotactic scores).

For comparison, examples of the names in the highest scored ('worst') group are shown below (full list shown in Appendix B).

Word type 664 - jajew (8.94), jejir (2.71), jojey (7.70), jyjar (5.84)

How do these groupings 'map' to those sets of names independently considered to be credible examples for brandability? One approach to answering this question involves looking at the domains offered for sale on brandable name marketplaces, on which the names have already been reviewed for suitability and deemed credible. In this case, I consider the top 2,000 5-character (SLD) names offered for sale on the domain marketplace Atom.com^[5] (filtering the results by requesting the inclusion only of 'made-up' names (since dictionary terms will presumably automatically be deemed more attractive, regardless of their phonetic characteristics), names up to two syllables, and sorting the results by the 'most popular' - though in this case, additional factors, such as domain asking-price, may become a factor).

Of these 2,000 names, 390 match the CVCVC pattern considered in this study. Table 2 shows the top ten word types represented within this dataset. Note that these most popular brand name word types are not, in general, those which were found to be most highly ranked (according to their mean phonotactic score) in the previous analysis.

Word type	No. domains	Word type ranking in phonotactic study (/ 216)
133	21	91
334	21	149
313	18	100
331	18	129
335	16	82
113	15	89
332	14	79
333	14	103
131	14	101
312	12	40

Table 2: Top ten word types represented in the dataset of most popular CVCVC names on the Atom.com domain marketplace

Again for illustration, the SLDs of the 21 domains assigned to the most common word-type group (for the set of ‘popular’ domains on Atom.com considered in this analysis) are listed below (again with their phonotactic scores shown in brackets).

Word type 133:

casux (0.00)
qafus (3.57)
qovox (3.70)
buvas (3.57)
buxas (1.05)
dosox (-)
pevox (-)
cuvax (-)
buvax (-)
koxas (0.41)
tevav (-)
kisiv (1.08)
qovax (5.19)
cuvux (14.73)
qasus (1.05)
qovas (1.55)
dezox (-)
pizah (2.04)
kavux (5.38)
bivoz (7.78)
tyvix (1.55)

Again, we see a lack of correlation of (just) phonotactic score with the subjective measure of desirability denoted by the inclusion of the names on the marketplace; the mean phonotactic violation score across the set of 390 CVCVC names taken from Atom.com is actually relatively high (3.36) (see Appendix C). Of course, it is a matter of opinion how 'good' these names 'actually' are, compared with any of the other candidates identified in these studies.

Discussion

What does all this tell us? One key take-away is that phonotactic scores in isolation do not provide a very good measure of the credibility of a domain name as a candidate for brandability (assuming that there is a possibility for any sort of non-subjective measure of desirability!), but that other simple groupings of names based on high-level phonetic characteristics (e.g. 'word types') of the strings in question can provide a basis for filtering candidate names for review.

The framework presented here is an extremely simplistic one, and it is likely that enhancements will be necessary in order to be able to generalise the approach at all - to take account, for example, of strings of differering lengths, alternative consonant / vowel patterns, inclusion of diphthongs or longer character combinations, or variability in the pronunciation of individual characters (such as treating the 'c' in 'ca*', 'co*' and 'cu*' (generally hard) differently from that in 'ce*' and 'ci*' (generally soft)).

Nevertheless, this type of approach does provide the potential for specifying the 'type' of word which may be of interest for a potential brand name, and using this specification as a basis for filtering the huge set of unregistered names which are to be considered. It is fair to say that we are still some way from being able to say that certain word types make 'better' brandable candidates, but it may be the case that extensive review of datasets by branding and marketing experts may help us move in this direction. If - for instance - we assume that word type 133 is a 'good' one for CVCVC names to serve as brandable candidates, based on the examples listed on the Atom.com marketplace, this would provide some basis for starting to review the huge set of unregistered .com names from the original study. There are 5,118 type-133 names in the original dataset (out of the 'pool' of over 137,000 CVCVC names), and some of these will undoubtedly be credible candidates for brandability. Exactly which ones will be a subjective matter of opinion, but examples might include names such as cisyx, civyz, cyfax, cyxix, doxiz, dyxix, gyfex, kysix, pyxox, qaxix, qizox, qyxes and toxiz.

* * * * *

Appendix A: Phonotactic analysis of the groupings of the unregistered CVCVC .com domains from the original study

Word type	No. domains (N)	Mean phonotactic violation score of domains in group	Ranking
225	15	0.41	1
525	4	0.50	2
555	4	0.75	3
125	84	0.82	4
222	23	0.92	5
425	39	0.98	6
512	85	0.99	7
122	175	1.14	8
215	77	1.19	9
422	73	1.27	10
625	12	1.35	11
255	12	1.36	12
155	85	1.38	13
322	125	1.39	14
252	4	1.43	15
325	47	1.44	16
622	28	1.45	17
522	14	1.51	18
115	577	1.65	19
152	49	1.66	20
212	158	1.67	21
265	15	1.76	22
315	407	1.82	23
415	226	1.92	24
352	18	1.98	25
112	1,074	2.00	26
615	84	2.00	27
235	43	2.08	28
612	161	2.09	29
355	57	2.09	30
565	14	2.12	31
412	401	2.16	32
552	5	2.18	33
221	225	2.20	34
655	13	2.21	35
551	46	2.27	36
251	85	2.28	37
455	41	2.32	38
515	41	2.33	39
312	792	2.33	40
553	28	2.37	41
532	48	2.37	42
135	427	2.38	43
151	556	2.40	44
521	130	2.41	45
121	1,245	2.54	46
153	400	2.54	47
132	752	2.55	48
452	26	2.59	49
165	129	2.65	50
513	534	2.70	51
451	257	2.71	52
453	152	2.75	53
123	955	2.78	54
351	412	2.79	55
523	80	2.82	56
254	49	2.82	57
421	532	2.83	58
432	281	2.84	59
465	55	2.89	60
511	760	2.89	61
232	93	2.94	62
223	158	2.96	63
262	26	2.99	64
621	176	3.00	65
665	15	3.03	66
154	218	3.05	67
253	62	3.11	68
213	1,122	3.14	69
632	105	3.17	70
353	284	3.19	71
365	109	3.22	72
321	992	3.22	73
535	31	3.23	74
531	592	3.24	75
435	162	3.29	76
424	192	3.29	77
423	383	3.29	78
332	565	3.30	79
323	710	3.31	80
354	182	3.32	81
335	308	3.32	82
211	1,434	3.37	83
226	70	3.40	84
554	26	3.47	85
454	91	3.47	86
456	71	3.48	87
533	445	3.51	88
113	5,988	3.51	89
562	14	3.53	90
133	5,118	3.53	91
233	853	3.55	92
145	241	3.56	93
124	518	3.58	94
623	118	3.60	95
514	279	3.60	96
413	2,376	3.62	97
652	7	3.65	98
645	32	3.66	99
313	4,733	3.69	100
131	6,740	3.69	101
111	7,833	3.73	102
333	3,805	3.76	103
433	1,872	3.81	104
245	41	3.85	105
524	50	3.85	106
411	3,225	3.87	107
114	2,832	3.87	108
224	96	3.90	109
414	1,104	3.90	110
231	1,140	3.91	111
324	409	3.92	112
266	51	3.93	113
314	2,297	3.94	114
126	332	3.94	115
633	709	3.95	116
311	6,734	3.95	117
662	23	3.96	118
613	887	3.96	119
431	2,607	3.96	120
214	575	3.98	121
635	47	3.98	122
142	383	3.99	123
242	57	4.01	124
542	26	4.03	125
442	163	4.05	126
566	30	4.05	127
534	211	4.05	128
331	5,396	4.12	129
611	1,202	4.15	130
563	128	4.27	131
426	140	4.27	132
614	388	4.28	133
166	248	4.30	134
541	324	4.31	135
626	37	4.33	136
631	954	4.33	137
236	296	4.34	138
134	2,404	4.36	139
216	392	4.37	140
653	63	4.37	141
445	103	4.39	142
466	88	4.42	143
326	269	4.48	144
545	20	4.48	145
162	197	4.49	146
141	3,052	4.50	147
256	31	4.57	148
334	1,808	4.59	149
116	1,697	4.61	150
642	59	4.61	151
263	206	4.62	152
651	85	4.62	153
416	698	4.64	154
556	14	4.65	155
516	188	4.66	156
345	195	4.66	157
434	835	4.69	158
234	435	4.69	159
342	324	4.69	160
241	613	4.71	161
624	69	4.72	162
616	213	4.72	163
543	246	4.76	164
163	1,060	4.76	165
436	534	4.77	166
462	83	4.79	167
441	1,221	4.83	168
156	153	4.83	169
264	87	4.91	170
356	130	4.92	171
663	132	4.93	172
143	2,267	4.94	173
362	165	4.96	174
536	140	4.96	175
136	1,442	5.01	176
636	167	5.09	177
463	396	5.09	178
316	1,399	5.09	179
443	884	5.13	180
341	2,557	5.14	181
526	32	5.19	182
641	426	5.20	183
161	1,459	5.20	184
366	173	5.21	185
654	35	5.24	186
661	168	5.29	187
643	307	5.30	188
634	314	5.38	189
164	467	5.40	190
343	1,765	5.48	191
561	162	5.50	192
363	805	5.52	193
261	270	5.59	194
243	440	5.62	195
461	560	5.65	196
336	1,060	5.67	197
546	70	5.77	198
244	245	5.80	199
446	249	5.92	200
144	1,146	5.93	201
564	49	5.96	202
656	20	5.98	203
146	613	5.99	204
444	430	6.01	205
646	77	6.02	206
361	1,219	6.12	207
464	161	6.15	208
364	373	6.23	209
344	943	6.45	210
666	23	6.52	211
346	456	6.61	212
544	122	6.92	213
644	158	6.96	214
246	140	7.00	215
664	53	7.15	216

Appendix B: Groups of individual domain SLDs by word type

a. Top four groups by (lowest) mean phonotactic violation score (shown in brackets in each case)

Word type 225 (total count, N = 15, mean phonotactic violation score, P̅ = 0.41):

memyl (0.00)
munyl (1.63)
nemul (0.00)
nenyl (0.00)
nimyl (1.05)
nomyl (1.05)
nonel (0.41)
nunyl (0.93)
nymal (1.05)
nymel (0.00)
nymul (0.00)
nymyl (0.00)
nynil (0.00)
nynul (0.00)
nynyl (0.00)

Word type 525 (N = 4, P̅ = 0.50):

lemyl (0.00)
limyl (1.05)
lunyl (0.93)
lynul (0.00)

Word type 555 (N = 4, P̅ = 0.75):

lalyl (0.90)
lelyl (0.00)
lylel (1.05)
lylul (1.05)

Word type 125 (N = 84, P̅ = 0.82):

banyl (0.12)
bemyl (0.12)
bimyl (1.17)
bomyl (0.52)
bumyl (1.05)
bymyl (1.17)
bynyl (1.17)
cemyl (0.00)
cenyl (0.00)
cimul (1.58)
cimyl (0.00)
cinul (0.00)
cunyl (1.63)
cynul (0.00)
cynyl (0.00)
dimyl (0.00)
dumyl (0.93)
dunyl (0.93)
dymul (0.00)
dymyl (0.00)
dynul (0.00)
gamyl (0.70)
ganyl (0.70)
gimyl (0.70)
gonyl (1.75)
gumyl (1.64)
gunyl (1.64)
gymol (7.62)
gynil (0.79)
gynul (0.79)
gynyl (0.79)
kumyl (0.93)
kunyl (0.93)
kymul (0.58)
kynal (1.05)
kynul (0.00)
kynyl (0.00)
pemyl (0.00)
pimyl (0.00)
pumyl (1.63)
punul (1.63)
punyl (1.63)
pymal (1.05)
pymul (0.00)
pymyl (1.05)
pynul (0.00)
pynyl (0.00)
qamyl (0.00)
qanul (0.00)
qanyl (0.58)
qemil (0.00)
qemyl (0.00)
qenal (2.10)
qenil (2.10)
qenul (0.00)
qenyl (0.00)
qimul (0.58)
qimyl (0.00)
qinel (0.00)
qinul (0.00)
qomil (0.41)
qomyl (0.41)
qonol (6.25)
qonul (1.38)
qonyl (1.05)
qumyl (1.91)
qunyl (1.91)
qymal (4.33)
qymil (0.97)
qymol (0.00)
qymul (1.63)
qynal (1.05)
qynel (0.00)
qynil (0.00)
qynol (0.00)
qynul (0.00)
qynyl (0.00)
temyl (0.00)
tumyl (0.93)
tunyl (0.93)
tymul (0.00)
tynil (1.05)
tynul (0.00)
tynyl (1.05)

b. Bottom group by (highest) mean phonotactic violation score (shown in brackets in each case)

Word type 664 (N = 53, P̅ = 7.15):

jajew (8.94)
jajiw (8.94)
jajiy (3.53)
jajur (7.28)
jajuy (6.72)
jajyr (3.76)
jajyw (7.11)
jejaw (8.94)
jejir (2.71)
jejiw (8.94)
jejiy (2.48)
jejow (8.94)
jejuw (8.94)
jejyr (2.97)
jejyw (-)
jejyy (2.48)
jijaw (16.25)
jijey (2.48)
jijiw (10.46)
jijuy (3.53)
jijyr (5.05)
jijyw (10.46)
jijyy (2.48)
jojaw (9.01)
jojew (9.99)
jojey (7.70)
jojir (12.79)
jojiw (7.11)
jojiy (7.70)
jojuw (9.99)
jojyr (6.92)
jojyw (7.11)
jujaw (9.99)
jujir (3.76)
jujiy (3.53)
jujow (4.54)
jujyr (3.76)
jujyw (7.11)
jujyy (3.53)
jyjar (5.84)
jyjaw (17.62)
jyjay (6.50)
jyjer (5.05)
jyjey (5.61)
jyjir (5.05)
jyjiw (22.87)
jyjiy (5.61)
jyjor (5.84)
jyjow (9.50)
jyjur (5.84)
jyjuw (12.07)
jyjuy (8.80)
jyjyy (5.61)

Appendix C: Set of popular CVCVC (SLD) names listed on the Atom.com marketplace (in order of listing on the site)

SLD name	Word type	Phonotactic violation score
wozot	431	6.66
vupor	314	-
vumol	325	5.72
buvit	131	1.55
ruvor	434	2.68
buvic	131	2.24
qoryx	143	1.05
lokem	512	1.05
wixoz	433	0.97
nuwix	243	8.48
modoq	211	8.49
casux	133	0.00
nenen	222	0.00
zivay	334	-
ryzix	433	0.99
sazol	335	1.40
hoxal	335	0.96
rusax	433	-
kutey	114	2.53
vyval	335	2.05
wodaz	413	3.35
veqan	312	1.55
vuvur	334	3.18
risox	433	-
zuzaz	333	-
fudox	313	5.06
qafus	133	3.57
vozom	332	2.54
wunax	423	-
voyiv	343	10.46
noyox	243	-
roloz	453	8.22
hypiq	311	3.76
bijil	165	19.24
wixah	433	0.97
rivun	432	0.50
vymec	321	5.19
qovox	133	3.70
zazoc	331	-
zovun	332	2.54
zifil	335	1.36
povor	134	2.68
jabus	613	0.90
kuboc	111	-
vavur	334	2.13
jovur	634	3.26
fivox	333	-
pexet	131	0.00
vitym	312	5.65
jovyx	633	2.33
jopex	613	5.48
rydat	411	1.05
juxan	632	1.72
ryzel	435	2.04
kaxil	135	0.00
tevam	132	-
magik	211	0.79
suruk	341	2.73
kutox	113	4.69
pivoy	134	-
zonoh	323	1.40
ruxic	431	0.93
joxar	634	-
ryzid	431	1.48
tuzic	131	2.04
puroq	141	6.01
favun	332	4.07
meyex	243	7.24
wyzal	435	2.54
vytul	315	1.55
huvah	333	1.99
humux	323	2.65
xaxas	333	0.99
kozil	135	2.04
xejus	363	2.62
xaxor	334	2.12
yiket	411	5.27
ryzil	435	0.99
pimox	123	-
vozaz	333	3.87
yaxum	432	0.58
hejix	363	1.35
zodem	312	2.04
dymox	123	4.69
povoy	134	6.24
ruxol	435	-
rykis	413	1.05
rynam	422	1.05
pivay	134	3.70
zybil	315	2.16
buvas	133	3.57
nohok	231	5.25
vuxel	335	1.43
jexas	633	0.79
zozic	331	7.53
yizic	431	2.62
vuxan	332	1.43
vybis	313	1.66
buxas	133	1.05
muvax	233	-
hoyox	343	11.40
jojur	664	7.93
vuzal	335	2.54
vuxar	334	2.56
nefik	231	0.37
nysor	234	2.18
zixar	334	2.12
cykal	115	1.05
yidox	413	-
dosox	133	-
wixas	433	0.97
vucin	312	1.55
zuvay	334	-
vacax	313	5.19
sevoz	333	0.50
moyox	243	5.00
fosyl	335	1.42
bomor	124	-
yador	414	2.76
pevox	133	-
jixol	635	0.79
jojex	663	-
zavem	332	18.44
vuxic	331	1.43
nyzon	232	0.99
koxol	135	3.11
nidom	212	0.00
vupax	313	5.19
vudit	311	1.55
fisil	335	0.37
zuzur	334	4.17
vijix	363	1.29
vevev	333	1.00
vudom	312	5.72
vybix	313	1.66
kykys	113	4.16
hoxor	334	2.09
vopom	312	1.55
vamoy	324	1.40
wimox	423	-
zovoy	334	7.23
juwab	641	5.49
nyzol	235	7.82
yaxam	432	0.58
kadoy	114	-
vusim	332	6.72
wizun	432	1.96
huxic	331	2.48
vybem	312	20.00
zuzat	331	-
duzar	134	3.17
vayox	343	7.73
hydux	313	1.61
vudix	313	1.55
gozok	131	-
pelas	153	0.00
yaxoy	434	3.28
wydus	413	0.97
zozet	331	3.03
temux	123	0.00
tyvix	133	1.55
zybal	315	2.16
fovay	334	-
vutay	314	-
qorik	141	5.00
vubos	313	13.01
xeror	344	-
vybon	312	5.31
xudor	314	3.17
fudax	313	11.83
zuxum	332	1.93
wyzox	433	-
yaxer	434	1.71
cuvax	133	-
vozay	334	3.44
cuzoc	131	-
zoxum	332	1.40
vabax	313	5.31
zixam	332	0.99
fofal	335	11.62
fivoy	334	1.77
buvax	133	-
xumox	323	5.69
vudas	313	2.88
modos	213	8.22
vofic	331	1.27
vutol	315	5.19
joxol	635	7.03
yutox	413	5.27
vyxon	332	2.52
vevup	331	2.26
zodum	312	2.04
wevum	432	1.47
jucid	611	2.33
dydox	113	5.78
pumol	125	1.05
fuvun	332	1.92
muvay	234	3.02
vudic	311	1.55
poyax	143	-
noler	254	2.18
zynol	325	2.04
vufic	331	6.10
zixol	335	0.99
viviy	334	1.90
koxas	133	0.41
vobac	311	1.66
vudec	311	5.72
tayox	143	7.24
zilyx	353	0.99
luyul	545	2.92
vaved	331	9.38
vupel	315	1.55
vypax	313	5.19
vudax	313	27.16
royaz	443	6.33
duryx	143	2.73
vudol	315	1.55
vapux	313	0.50
zinoc	321	5.69
zovyk	331	2.54
voyiz	343	9.86
fetiz	313	0.37
qerol	145	0.00
koxil	135	0.41
lovyz	533	2.88
loyix	543	5.00
bybix	113	1.28
bofed	131	6.33
vurin	342	3.23
vazen	332	2.54
hyqor	314	5.25
wyzow	434	4.69
totix	113	1.05
doxay	134	-
lixyl	535	0.00
ryzim	432	0.99
watif	413	3.43
tevav	133	-
xevix	333	1.49
qidix	113	0.00
vuvec	331	-
nipod	211	0.49
zayax	343	3.91
lavun	532	3.70
qonor	124	1.54
pozen	132	2.04
sosel	335	1.05
vazyl	335	2.54
wilic	451	2.02
balac	151	0.52
coboh	113	3.87
nomel	225	1.05
kejat	161	2.89
sohor	334	7.23
guvit	131	2.14
muzuh	233	5.23
juvyc	631	2.33
vavov	333	2.08
zojal	365	2.83
quliv	153	4.25
hixir	334	1.69
wepuf	413	3.03
nuvux	233	1.55
zidux	313	0.99
zylon	352	2.04
sesef	333	2.06
zupak	311	2.04
noboh	213	3.87
voras	343	1.55
femez	323	-
sinez	323	-
zulec	351	2.04
fevum	332	1.92
vodiq	311	1.55
widoz	413	8.14
mytes	213	3.90
yariq	441	0.58
tupux	113	10.48
lyzic	531	2.04
mogux	213	1.75
juvyn	632	2.22
kisiv	133	1.08
wiqet	411	0.97
qovax	133	5.19
wipux	413	0.97
zojon	362	6.47
jurux	643	9.84
xalix	353	0.99
hygic	311	2.40
hilis	353	0.56
sazay	334	-
zotey	314	2.94
toruq	141	1.05
vudem	312	1.55
vysal	335	1.55
pypic	111	1.05
qical	115	1.05
cuvux	133	14.73
wavat	431	1.88
deqox	113	5.96
vefut	331	0.87
zoyac	341	6.00
vusar	334	2.68
pacol	115	1.05
lyxar	534	1.13
ciniv	123	1.08
kivoc	131	0.50
wisox	433	-
gavum	132	1.20
vubux	313	5.84
fivay	334	6.61
zazay	334	-
tucus	113	1.05
vopax	313	5.19
miqos	213	6.89
tixot	131	0.00
qobon	112	3.32
voyab	341	7.28
zeyox	343	8.23
zizot	331	1.98
fexen	332	0.37
nuvam	232	-
zatam	312	2.04
zizet	331	3.91
dixel	135	0.00
cabaz	113	-
zanic	321	0.99
begax	113	-
nibax	213	4.81
vufex	333	5.56
qasus	133	1.05
ruvam	432	-
budax	113	11.67
jurer	644	4.65
vezil	335	1.49
fuqet	311	5.71
qovas	133	1.55
zaroc	341	-
fikay	314	6.11
wajex	463	6.45
cuned	121	2.11
menos	223	8.22
dogoz	113	8.28
dezox	133	-
nozov	233	-
pizah	133	2.04
tuyul	145	5.00
pikav	113	-
duxul	135	0.93
vuxos	333	7.27
wuxer	434	3.04
teqer	114	2.18
neloz	253	0.00
givoc	131	1.20
bexat	131	-
nesux	233	0.00
vuvul	335	2.05
nyvix	233	1.55
povem	132	-
hejex	363	1.35
jemox	623	6.56
vovav	333	-
moxur	234	1.54
hoxed	331	1.78
qanor	124	1.13
vudur	314	2.68
movym	232	11.57
vojov	363	2.05
jymex	623	6.56
qalon	152	2.56
lydax	513	4.69
bewat	141	5.47
cesot	131	-
zunek	321	2.04
huxey	334	5.03
yerey	444	1.48
hozet	331	2.60
kavux	133	5.38
cozok	131	5.69
fumoc	321	5.64
vypon	312	1.55
tycer	114	2.18
zuvil	335	2.54
ryzus	433	0.99
bivoz	133	7.78
hosox	333	-
qyler	154	2.18
jeloc	651	0.79
fasyt	331	1.42