Another 802.11n question

Edit:  to add more context on the neighbor nodes.

Note, that under OpenWRT Barrier Breaker (that 3.16.1.0b1 is on top of), only the 802.11n LGI (Long Guard Interval at 800ns) is supported.  Once we upgrade to Chaos Calmer, this will include options to use the SGI (Short Guard Interval at 400 ns) = a  littler higher rates.   Consequently, make sure you are looking at the correct rate tables.  

An 802.11n node, when communicating with another 802.11n neighbor, builds a rate selection table and tests the various rate options (<10% searching around of packets).  It keeps statistics of the success rates and picks the maximum throughput option (which is show now in mesh status).  The rate selection table looks like this.  These tables are all basd on 20Mhz channel width.   Big "T" is best throughput, little "t' is 2nd best.    This neighbor is ~10 miles away--a RocketDish.

type         rate     throughput  ewma prob   this prob  retry   this succ/attempt   success    attempts
HT20/LGI     MCS0            5.6      100.0       100.0      1              0(  0)         1           1
HT20/LGI     MCS1           10.5      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS2           14.8      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS3           18.6      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS4           25.1      100.0       100.0      5              0(  0)         1           1
HT20/LGI     MCS5           30.3       98.1       100.0      5              0(  0)        48          51
HT20/LGI   P MCS6           32.5       92.7       100.0      5              0(  0)        48          69
HT20/LGI     MCS7           31.8       81.8       100.0      5              0(  0)        35          70
HT20/LGI     MCS8           10.5      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS9           18.6      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS10          25.1      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS11          30.3       99.9       100.0      5              0(  0)        32          33
HT20/LGI  t  MCS12          37.6       87.4        50.0      6              0(  0)      4681        4981
HT20/LGI T   MCS13          44.0       97.8       100.0      6              0(  0)      4142        4917
HT20/LGI     MCS14          16.3       31.6       100.0      6              0(  0)       412         710
HT20/LGI     MCS15           0.0        1.4         0.0      6              0(  0)        11         132

Total packet count::    ideal 9082      lookaround 336
Average A-MPDU length: 1.0

Note, if the neighbor was only a single antenna device (non-MIMO), then the MCS8 to MCS15 rates would not show (which are 2 spatial streams of data).

When the same 802.11n node talks to another device (at the same time) that is only 802.11a/g, the rate selection it looks to send at matches.  "A" is best throughput, "B" is next best, etc.   This neighbor is 8 miles away, is a NanoBridge, and not yet upgraded to 802.11n.  This is what it looked like historically with BBHN/AREDN in all releases.

Average A-MPDU length: 1.0
rate      throughput  ewma prob  this prob  this succ/attempt   success    attempts
       6         1.5       25.0      100.0             0(  0)         1           1
       9         2.2       25.0      100.0             0(  0)         1           1
      12         2.9       25.0      100.0             0(  0)         1           1
      18         4.3       25.0      100.0             0(  0)         1           1
   D  24         5.6       25.0      100.0             0(  0)         1           1
  C   36         8.1       25.0      100.0             0(  0)         1           1
A   P 48        41.5       99.7      100.0             1(  1)       150         155
 B    54        20.2       43.7      100.0             0(  0)         2           2

Total packet count::    ideal 142      lookaround 16

To compare the throughput between these specific rate tables, there may not be enough packet attempts to have a good characterization of the link--150 packets is probably not enough to characterize the link.  Best send HD video over the link first :) .