What is a good reject rate during rapid blocks?

[dbbit]

I am solo mining some SHA256 coins that sometimes changes blocks rapidly (<10 seconds). I have a pretty high reject rate (20%) which always seems to do with finding a block within a few seconds after a switch.

Even with the 20% reject rate, I'm still doing better on average than I would do in a pool though.  Is 20% rejects just expected in this case? Is that what the pools are also facing? And most importantly - is there anything I can do to make it better?

* Mining gear runs at ~420gh/s
* Connecting over 100mps to do GetWork against a -q wallet.
* Internet connection is 60mbps up, 10mbps down.
* Queue size is 30. If I make it lower than 30, I see a lot of "Pool x not providing work fast enough". If I make it higher I can see it submits blocks 5 or 10 seconds after change. Not good.


--------------------------------------------------------------------------------                                                
 (5s):452.2G (avg):437.7Gh/s | A:977635  R:250404  HW:2493  WU:7946.5/m                                                        
 ST: 27  SS: 0  NB: 231  LW: 340449  GF: 65  RF: 1                                                                              
 Connected to 192.168.2.13 diff 20.5K with LP as user xxxx (cgminer is confused with LP flag. I have a failover pool that uses LP. Active pool doesn't have Longpoll.)
 Block: 11569b68...  Diff:49.5K  Started: [23:56:42]  Best share: 971K                                                        
--------------------------------------------------------------------------------        
 HFA 0: 424.6G/438.0Gh/s | A:977635 R:250404 HW:2502 WU:7937.3/m                            
--------------------------------------------------------------------------------  
 [2014-02-21 23:46:44] Pool 0 not providing work fast enough    
 [2014-02-23 23:46:47] Network diff set to 49.5K                                                                                
 [2014-02-23 23:46:47] New block detected on network before longpoll                                                            
 [2014-02-23 23:46:53] New block detected on network before longpoll                                                            
 [2014-02-23 23:46:55] Found block for pool 0!                                                                                  
 [2014-02-23 23:46:55] Pool 0 stale share detected, submitting as user requested                                                
 [2014-02-23 23:46:55] Rejected 11471acf Diff 971K/49527 BLOCK! HFA 0 pool 0                                                    
 [2014-02-23 23:46:56] New block detected on network before longpoll                                                            
 [2014-02-23 23:46:56] Pool 0 not providing work fast enough                                                                    

[cescan]

Some gpu just suck at mining.

[dbbit]

Some gpu just suck at mining.

You know of many GPU's that can get 420 gh/s doing SHA256?

It's an ASIC, not a GPU.